{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 8.0,
  "eval_steps": 500,
  "global_step": 29352,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "entropy": 1.0396083772182465,
      "epoch": 0.002725538293813028,
      "grad_norm": 5.405938148498535,
      "learning_rate": 3.2703488372093024e-07,
      "loss": 1.8885,
      "mean_token_accuracy": 0.6852266788482666,
      "num_tokens": 711886.0,
      "step": 10
    },
    {
      "entropy": 1.0463788136839867,
      "epoch": 0.005451076587626056,
      "grad_norm": 5.517510890960693,
      "learning_rate": 6.904069767441861e-07,
      "loss": 1.8765,
      "mean_token_accuracy": 0.680418786406517,
      "num_tokens": 1416062.0,
      "step": 20
    },
    {
      "entropy": 1.0476338595151902,
      "epoch": 0.008176614881439084,
      "grad_norm": 1.9664850234985352,
      "learning_rate": 1.0537790697674419e-06,
      "loss": 1.7861,
      "mean_token_accuracy": 0.6949809923768043,
      "num_tokens": 2149938.0,
      "step": 30
    },
    {
      "entropy": 1.0441522210836411,
      "epoch": 0.010902153175252113,
      "grad_norm": 0.7536147832870483,
      "learning_rate": 1.4171511627906977e-06,
      "loss": 1.7475,
      "mean_token_accuracy": 0.6960390791296959,
      "num_tokens": 2859500.0,
      "step": 40
    },
    {
      "entropy": 1.020177149772644,
      "epoch": 0.01362769146906514,
      "grad_norm": 0.7229697108268738,
      "learning_rate": 1.7805232558139537e-06,
      "loss": 1.6856,
      "mean_token_accuracy": 0.7006200030446053,
      "num_tokens": 3568762.0,
      "step": 50
    },
    {
      "entropy": 1.0083495900034904,
      "epoch": 0.016353229762878167,
      "grad_norm": 0.5624799728393555,
      "learning_rate": 2.1438953488372095e-06,
      "loss": 1.6417,
      "mean_token_accuracy": 0.7085390493273735,
      "num_tokens": 4288749.0,
      "step": 60
    },
    {
      "entropy": 1.0113003820180893,
      "epoch": 0.019078768056691196,
      "grad_norm": 0.4210813045501709,
      "learning_rate": 2.5072674418604653e-06,
      "loss": 1.6123,
      "mean_token_accuracy": 0.7114065825939179,
      "num_tokens": 5021234.0,
      "step": 70
    },
    {
      "entropy": 1.0052868500351906,
      "epoch": 0.021804306350504225,
      "grad_norm": 0.642890214920044,
      "learning_rate": 2.870639534883721e-06,
      "loss": 1.5838,
      "mean_token_accuracy": 0.7120683744549752,
      "num_tokens": 5739006.0,
      "step": 80
    },
    {
      "entropy": 1.009394559264183,
      "epoch": 0.024529844644317254,
      "grad_norm": 0.47518935799598694,
      "learning_rate": 3.2340116279069773e-06,
      "loss": 1.5737,
      "mean_token_accuracy": 0.722443425655365,
      "num_tokens": 6477781.0,
      "step": 90
    },
    {
      "entropy": 1.0062545910477638,
      "epoch": 0.02725538293813028,
      "grad_norm": 0.5360950231552124,
      "learning_rate": 3.597383720930233e-06,
      "loss": 1.5498,
      "mean_token_accuracy": 0.7235137611627579,
      "num_tokens": 7213793.0,
      "step": 100
    },
    {
      "entropy": 1.0027928322553634,
      "epoch": 0.02998092123194331,
      "grad_norm": 0.7297307252883911,
      "learning_rate": 3.960755813953489e-06,
      "loss": 1.5363,
      "mean_token_accuracy": 0.7214894682168961,
      "num_tokens": 7936928.0,
      "step": 110
    },
    {
      "entropy": 1.0116635337471962,
      "epoch": 0.032706459525756335,
      "grad_norm": 0.5194182991981506,
      "learning_rate": 4.324127906976744e-06,
      "loss": 1.5155,
      "mean_token_accuracy": 0.7188240200281143,
      "num_tokens": 8650197.0,
      "step": 120
    },
    {
      "entropy": 1.0116185247898102,
      "epoch": 0.035431997819569364,
      "grad_norm": 0.6806616187095642,
      "learning_rate": 4.6875000000000004e-06,
      "loss": 1.5082,
      "mean_token_accuracy": 0.7230221316218376,
      "num_tokens": 9375831.0,
      "step": 130
    },
    {
      "entropy": 1.001607432961464,
      "epoch": 0.03815753611338239,
      "grad_norm": 0.6176725029945374,
      "learning_rate": 5.050872093023256e-06,
      "loss": 1.4882,
      "mean_token_accuracy": 0.7240102514624596,
      "num_tokens": 10093736.0,
      "step": 140
    },
    {
      "entropy": 1.0051208674907683,
      "epoch": 0.04088307440719542,
      "grad_norm": 0.6676029562950134,
      "learning_rate": 5.414244186046512e-06,
      "loss": 1.4921,
      "mean_token_accuracy": 0.7243235290050507,
      "num_tokens": 10817100.0,
      "step": 150
    },
    {
      "entropy": 1.0007067069411277,
      "epoch": 0.04360861270100845,
      "grad_norm": 0.578899621963501,
      "learning_rate": 5.777616279069767e-06,
      "loss": 1.4764,
      "mean_token_accuracy": 0.7260746091604233,
      "num_tokens": 11528126.0,
      "step": 160
    },
    {
      "entropy": 1.001201516389847,
      "epoch": 0.04633415099482148,
      "grad_norm": 0.5769525170326233,
      "learning_rate": 6.140988372093024e-06,
      "loss": 1.4712,
      "mean_token_accuracy": 0.7269342973828316,
      "num_tokens": 12260702.0,
      "step": 170
    },
    {
      "entropy": 1.0025376126170158,
      "epoch": 0.04905968928863451,
      "grad_norm": 0.6673255562782288,
      "learning_rate": 6.50436046511628e-06,
      "loss": 1.4798,
      "mean_token_accuracy": 0.7271475344896317,
      "num_tokens": 12983673.0,
      "step": 180
    },
    {
      "entropy": 0.9962176263332367,
      "epoch": 0.05178522758244753,
      "grad_norm": 0.8359505534172058,
      "learning_rate": 6.867732558139535e-06,
      "loss": 1.4674,
      "mean_token_accuracy": 0.725401483476162,
      "num_tokens": 13688416.0,
      "step": 190
    },
    {
      "entropy": 0.9942928344011307,
      "epoch": 0.05451076587626056,
      "grad_norm": 0.6369069814682007,
      "learning_rate": 7.231104651162791e-06,
      "loss": 1.4617,
      "mean_token_accuracy": 0.7302308842539788,
      "num_tokens": 14403522.0,
      "step": 200
    },
    {
      "entropy": 0.9998625546693802,
      "epoch": 0.05723630417007359,
      "grad_norm": 0.9715228080749512,
      "learning_rate": 7.594476744186047e-06,
      "loss": 1.4703,
      "mean_token_accuracy": 0.7334344819188118,
      "num_tokens": 15133763.0,
      "step": 210
    },
    {
      "entropy": 0.9997459784150123,
      "epoch": 0.05996184246388662,
      "grad_norm": 0.7882207036018372,
      "learning_rate": 7.957848837209303e-06,
      "loss": 1.462,
      "mean_token_accuracy": 0.727087727189064,
      "num_tokens": 15854077.0,
      "step": 220
    },
    {
      "entropy": 0.9903041675686837,
      "epoch": 0.06268738075769964,
      "grad_norm": 0.8622173070907593,
      "learning_rate": 8.321220930232558e-06,
      "loss": 1.461,
      "mean_token_accuracy": 0.7340906202793122,
      "num_tokens": 16577544.0,
      "step": 230
    },
    {
      "entropy": 1.0018442422151566,
      "epoch": 0.06541291905151267,
      "grad_norm": 0.8568950891494751,
      "learning_rate": 8.684593023255814e-06,
      "loss": 1.4658,
      "mean_token_accuracy": 0.7301117062568665,
      "num_tokens": 17310778.0,
      "step": 240
    },
    {
      "entropy": 0.9931416124105453,
      "epoch": 0.0681384573453257,
      "grad_norm": 0.7365680932998657,
      "learning_rate": 9.04796511627907e-06,
      "loss": 1.461,
      "mean_token_accuracy": 0.7339939594268798,
      "num_tokens": 18037717.0,
      "step": 250
    },
    {
      "entropy": 0.9978902578353882,
      "epoch": 0.07086399563913873,
      "grad_norm": 0.5008879899978638,
      "learning_rate": 9.411337209302326e-06,
      "loss": 1.4587,
      "mean_token_accuracy": 0.7276411607861519,
      "num_tokens": 18740674.0,
      "step": 260
    },
    {
      "entropy": 1.0006747156381608,
      "epoch": 0.07358953393295176,
      "grad_norm": 0.6403897404670715,
      "learning_rate": 9.77470930232558e-06,
      "loss": 1.4539,
      "mean_token_accuracy": 0.729505018889904,
      "num_tokens": 19469047.0,
      "step": 270
    },
    {
      "entropy": 0.999770636856556,
      "epoch": 0.07631507222676479,
      "grad_norm": 0.640429675579071,
      "learning_rate": 1.0138081395348837e-05,
      "loss": 1.4655,
      "mean_token_accuracy": 0.7282736241817475,
      "num_tokens": 20178240.0,
      "step": 280
    },
    {
      "entropy": 0.9961532056331635,
      "epoch": 0.07904061052057781,
      "grad_norm": 0.5266719460487366,
      "learning_rate": 1.0501453488372095e-05,
      "loss": 1.4558,
      "mean_token_accuracy": 0.7296340927481652,
      "num_tokens": 20889492.0,
      "step": 290
    },
    {
      "entropy": 0.9954295739531517,
      "epoch": 0.08176614881439084,
      "grad_norm": 0.8393326997756958,
      "learning_rate": 1.086482558139535e-05,
      "loss": 1.4485,
      "mean_token_accuracy": 0.7313675001263619,
      "num_tokens": 21612600.0,
      "step": 300
    },
    {
      "entropy": 0.9983824446797371,
      "epoch": 0.08449168710820387,
      "grad_norm": 0.5021800994873047,
      "learning_rate": 1.1228197674418606e-05,
      "loss": 1.456,
      "mean_token_accuracy": 0.7298679947853088,
      "num_tokens": 22331792.0,
      "step": 310
    },
    {
      "entropy": 0.9956601366400719,
      "epoch": 0.0872172254020169,
      "grad_norm": 0.48453429341316223,
      "learning_rate": 1.159156976744186e-05,
      "loss": 1.4574,
      "mean_token_accuracy": 0.7311487272381783,
      "num_tokens": 23042188.0,
      "step": 320
    },
    {
      "entropy": 0.9930527314543725,
      "epoch": 0.08994276369582993,
      "grad_norm": 0.5262903571128845,
      "learning_rate": 1.1954941860465118e-05,
      "loss": 1.4524,
      "mean_token_accuracy": 0.7352466434240341,
      "num_tokens": 23767525.0,
      "step": 330
    },
    {
      "entropy": 1.0016896530985833,
      "epoch": 0.09266830198964296,
      "grad_norm": 0.5496646761894226,
      "learning_rate": 1.2318313953488372e-05,
      "loss": 1.4541,
      "mean_token_accuracy": 0.7287753179669381,
      "num_tokens": 24484817.0,
      "step": 340
    },
    {
      "entropy": 0.9894623726606369,
      "epoch": 0.09539384028345599,
      "grad_norm": 0.4871159791946411,
      "learning_rate": 1.2681686046511629e-05,
      "loss": 1.4469,
      "mean_token_accuracy": 0.7348430410027504,
      "num_tokens": 25215335.0,
      "step": 350
    },
    {
      "entropy": 0.9919575676321983,
      "epoch": 0.09811937857726902,
      "grad_norm": 0.5147045850753784,
      "learning_rate": 1.3045058139534885e-05,
      "loss": 1.4484,
      "mean_token_accuracy": 0.7358337089419364,
      "num_tokens": 25947042.0,
      "step": 360
    },
    {
      "entropy": 0.997435276210308,
      "epoch": 0.10084491687108203,
      "grad_norm": 0.5011497139930725,
      "learning_rate": 1.340843023255814e-05,
      "loss": 1.4594,
      "mean_token_accuracy": 0.7311121672391891,
      "num_tokens": 26657723.0,
      "step": 370
    },
    {
      "entropy": 0.9983295291662216,
      "epoch": 0.10357045516489506,
      "grad_norm": 0.7181002497673035,
      "learning_rate": 1.3771802325581396e-05,
      "loss": 1.4636,
      "mean_token_accuracy": 0.7269156962633133,
      "num_tokens": 27371239.0,
      "step": 380
    },
    {
      "entropy": 1.0010447755455971,
      "epoch": 0.10629599345870809,
      "grad_norm": 0.5289329290390015,
      "learning_rate": 1.4135174418604652e-05,
      "loss": 1.4527,
      "mean_token_accuracy": 0.7297766819596291,
      "num_tokens": 28080993.0,
      "step": 390
    },
    {
      "entropy": 1.0036471426486968,
      "epoch": 0.10902153175252112,
      "grad_norm": 0.5319470167160034,
      "learning_rate": 1.4498546511627908e-05,
      "loss": 1.4658,
      "mean_token_accuracy": 0.7272291526198387,
      "num_tokens": 28788262.0,
      "step": 400
    },
    {
      "entropy": 0.9968745350837708,
      "epoch": 0.11174707004633415,
      "grad_norm": 0.9511128067970276,
      "learning_rate": 1.4861918604651163e-05,
      "loss": 1.4558,
      "mean_token_accuracy": 0.7349012956023216,
      "num_tokens": 29521970.0,
      "step": 410
    },
    {
      "entropy": 0.9982865825295448,
      "epoch": 0.11447260834014718,
      "grad_norm": 0.44772422313690186,
      "learning_rate": 1.5225290697674419e-05,
      "loss": 1.4502,
      "mean_token_accuracy": 0.7337453439831734,
      "num_tokens": 30246421.0,
      "step": 420
    },
    {
      "entropy": 1.0040655449032783,
      "epoch": 0.1171981466339602,
      "grad_norm": 0.41664543747901917,
      "learning_rate": 1.5588662790697675e-05,
      "loss": 1.4619,
      "mean_token_accuracy": 0.7275902599096298,
      "num_tokens": 30948758.0,
      "step": 430
    },
    {
      "entropy": 0.9926177501678467,
      "epoch": 0.11992368492777324,
      "grad_norm": 0.4886620044708252,
      "learning_rate": 1.595203488372093e-05,
      "loss": 1.4539,
      "mean_token_accuracy": 0.7322545900940896,
      "num_tokens": 31651728.0,
      "step": 440
    },
    {
      "entropy": 0.9997780650854111,
      "epoch": 0.12264922322158626,
      "grad_norm": 0.4385494887828827,
      "learning_rate": 1.6315406976744187e-05,
      "loss": 1.4608,
      "mean_token_accuracy": 0.7317872077226639,
      "num_tokens": 32363433.0,
      "step": 450
    },
    {
      "entropy": 0.9970521241426468,
      "epoch": 0.12537476151539928,
      "grad_norm": 0.42223232984542847,
      "learning_rate": 1.667877906976744e-05,
      "loss": 1.4553,
      "mean_token_accuracy": 0.7333403721451759,
      "num_tokens": 33077842.0,
      "step": 460
    },
    {
      "entropy": 1.00070518553257,
      "epoch": 0.1281002998092123,
      "grad_norm": 0.3606317639350891,
      "learning_rate": 1.70421511627907e-05,
      "loss": 1.4565,
      "mean_token_accuracy": 0.7299782425165177,
      "num_tokens": 33790314.0,
      "step": 470
    },
    {
      "entropy": 0.9955750495195389,
      "epoch": 0.13082583810302534,
      "grad_norm": 0.5059699416160583,
      "learning_rate": 1.7405523255813956e-05,
      "loss": 1.4453,
      "mean_token_accuracy": 0.7349987089633941,
      "num_tokens": 34520367.0,
      "step": 480
    },
    {
      "entropy": 0.9910548433661461,
      "epoch": 0.13355137639683837,
      "grad_norm": 0.3608299493789673,
      "learning_rate": 1.7768895348837212e-05,
      "loss": 1.4486,
      "mean_token_accuracy": 0.7293318182229995,
      "num_tokens": 35227706.0,
      "step": 490
    },
    {
      "entropy": 0.9997168734669686,
      "epoch": 0.1362769146906514,
      "grad_norm": 0.4888944923877716,
      "learning_rate": 1.8132267441860465e-05,
      "loss": 1.4487,
      "mean_token_accuracy": 0.7292981430888176,
      "num_tokens": 35943623.0,
      "step": 500
    },
    {
      "entropy": 0.9843390792608261,
      "epoch": 0.13900245298446443,
      "grad_norm": 0.44995641708374023,
      "learning_rate": 1.849563953488372e-05,
      "loss": 1.4392,
      "mean_token_accuracy": 0.7392399966716766,
      "num_tokens": 36675247.0,
      "step": 510
    },
    {
      "entropy": 0.9930981233716011,
      "epoch": 0.14172799127827745,
      "grad_norm": 0.6165155172348022,
      "learning_rate": 1.8859011627906978e-05,
      "loss": 1.4444,
      "mean_token_accuracy": 0.7339754477143288,
      "num_tokens": 37402108.0,
      "step": 520
    },
    {
      "entropy": 0.9964079529047012,
      "epoch": 0.14445352957209048,
      "grad_norm": 0.5399088859558105,
      "learning_rate": 1.9222383720930234e-05,
      "loss": 1.4563,
      "mean_token_accuracy": 0.7294779628515243,
      "num_tokens": 38109593.0,
      "step": 530
    },
    {
      "entropy": 0.984428557753563,
      "epoch": 0.1471790678659035,
      "grad_norm": 0.6215316653251648,
      "learning_rate": 1.9585755813953487e-05,
      "loss": 1.4321,
      "mean_token_accuracy": 0.737668377161026,
      "num_tokens": 38836506.0,
      "step": 540
    },
    {
      "entropy": 0.9932869285345077,
      "epoch": 0.14990460615971654,
      "grad_norm": 1.7897403240203857,
      "learning_rate": 1.9949127906976743e-05,
      "loss": 1.4469,
      "mean_token_accuracy": 0.7332193359732628,
      "num_tokens": 39552077.0,
      "step": 550
    },
    {
      "entropy": 0.9968496710062027,
      "epoch": 0.15263014445352957,
      "grad_norm": 0.7561520934104919,
      "learning_rate": 2.0312500000000002e-05,
      "loss": 1.4546,
      "mean_token_accuracy": 0.7331880688667297,
      "num_tokens": 40276845.0,
      "step": 560
    },
    {
      "entropy": 0.9941284239292145,
      "epoch": 0.1553556827473426,
      "grad_norm": 0.5177472829818726,
      "learning_rate": 2.067587209302326e-05,
      "loss": 1.4509,
      "mean_token_accuracy": 0.7373390078544617,
      "num_tokens": 40999687.0,
      "step": 570
    },
    {
      "entropy": 0.995505753159523,
      "epoch": 0.15808122104115563,
      "grad_norm": 0.384387731552124,
      "learning_rate": 2.1039244186046515e-05,
      "loss": 1.4522,
      "mean_token_accuracy": 0.7329975292086601,
      "num_tokens": 41709294.0,
      "step": 580
    },
    {
      "entropy": 0.9965997755527496,
      "epoch": 0.16080675933496866,
      "grad_norm": 0.4045581817626953,
      "learning_rate": 2.1402616279069768e-05,
      "loss": 1.4433,
      "mean_token_accuracy": 0.7339462488889694,
      "num_tokens": 42431630.0,
      "step": 590
    },
    {
      "entropy": 0.986753125488758,
      "epoch": 0.1635322976287817,
      "grad_norm": 0.3861728310585022,
      "learning_rate": 2.1765988372093024e-05,
      "loss": 1.4383,
      "mean_token_accuracy": 0.7369671985507011,
      "num_tokens": 43156812.0,
      "step": 600
    },
    {
      "entropy": 0.9907051742076873,
      "epoch": 0.16625783592259472,
      "grad_norm": 0.3145368695259094,
      "learning_rate": 2.212936046511628e-05,
      "loss": 1.4497,
      "mean_token_accuracy": 0.7377771839499474,
      "num_tokens": 43878714.0,
      "step": 610
    },
    {
      "entropy": 0.9945615753531456,
      "epoch": 0.16898337421640774,
      "grad_norm": 0.4484099745750427,
      "learning_rate": 2.2492732558139536e-05,
      "loss": 1.4465,
      "mean_token_accuracy": 0.732744038105011,
      "num_tokens": 44596974.0,
      "step": 620
    },
    {
      "entropy": 0.9958405807614327,
      "epoch": 0.17170891251022077,
      "grad_norm": 0.31601184606552124,
      "learning_rate": 2.285610465116279e-05,
      "loss": 1.4523,
      "mean_token_accuracy": 0.7347798213362694,
      "num_tokens": 45317391.0,
      "step": 630
    },
    {
      "entropy": 0.9867362692952156,
      "epoch": 0.1744344508040338,
      "grad_norm": 0.4932362139225006,
      "learning_rate": 2.321947674418605e-05,
      "loss": 1.4267,
      "mean_token_accuracy": 0.7379182651638985,
      "num_tokens": 46043618.0,
      "step": 640
    },
    {
      "entropy": 0.9920970141887665,
      "epoch": 0.17715998909784683,
      "grad_norm": 0.3708631098270416,
      "learning_rate": 2.3582848837209305e-05,
      "loss": 1.4486,
      "mean_token_accuracy": 0.7326340481638909,
      "num_tokens": 46759967.0,
      "step": 650
    },
    {
      "entropy": 0.989774090051651,
      "epoch": 0.17988552739165986,
      "grad_norm": 0.3645494878292084,
      "learning_rate": 2.394622093023256e-05,
      "loss": 1.4442,
      "mean_token_accuracy": 0.732268339395523,
      "num_tokens": 47473139.0,
      "step": 660
    },
    {
      "entropy": 0.9853468403220177,
      "epoch": 0.1826110656854729,
      "grad_norm": 0.32583683729171753,
      "learning_rate": 2.4309593023255814e-05,
      "loss": 1.4319,
      "mean_token_accuracy": 0.737530317902565,
      "num_tokens": 48201841.0,
      "step": 670
    },
    {
      "entropy": 0.9894756719470024,
      "epoch": 0.18533660397928592,
      "grad_norm": 0.3289868235588074,
      "learning_rate": 2.467296511627907e-05,
      "loss": 1.4426,
      "mean_token_accuracy": 0.7325257658958435,
      "num_tokens": 48908078.0,
      "step": 680
    },
    {
      "entropy": 0.98998514264822,
      "epoch": 0.18806214227309895,
      "grad_norm": 0.36332690715789795,
      "learning_rate": 2.5036337209302323e-05,
      "loss": 1.4361,
      "mean_token_accuracy": 0.7376216873526573,
      "num_tokens": 49638103.0,
      "step": 690
    },
    {
      "entropy": 0.9941335126757622,
      "epoch": 0.19078768056691198,
      "grad_norm": 0.34593939781188965,
      "learning_rate": 2.539970930232558e-05,
      "loss": 1.4457,
      "mean_token_accuracy": 0.7329727828502655,
      "num_tokens": 50354451.0,
      "step": 700
    },
    {
      "entropy": 0.9875242754817009,
      "epoch": 0.193513218860725,
      "grad_norm": 0.4945123493671417,
      "learning_rate": 2.5763081395348836e-05,
      "loss": 1.4405,
      "mean_token_accuracy": 0.7297769740223885,
      "num_tokens": 51051389.0,
      "step": 710
    },
    {
      "entropy": 0.9995409071445465,
      "epoch": 0.19623875715453803,
      "grad_norm": 0.38803181052207947,
      "learning_rate": 2.6126453488372095e-05,
      "loss": 1.4599,
      "mean_token_accuracy": 0.7296995922923089,
      "num_tokens": 51753445.0,
      "step": 720
    },
    {
      "entropy": 0.9887203484773636,
      "epoch": 0.19896429544835104,
      "grad_norm": 0.3487159013748169,
      "learning_rate": 2.648982558139535e-05,
      "loss": 1.447,
      "mean_token_accuracy": 0.732439212501049,
      "num_tokens": 52465342.0,
      "step": 730
    },
    {
      "entropy": 0.9897234991192818,
      "epoch": 0.20168983374216407,
      "grad_norm": 0.36995023488998413,
      "learning_rate": 2.6853197674418608e-05,
      "loss": 1.4405,
      "mean_token_accuracy": 0.732276639342308,
      "num_tokens": 53170427.0,
      "step": 740
    },
    {
      "entropy": 0.9915264055132866,
      "epoch": 0.2044153720359771,
      "grad_norm": 0.853056013584137,
      "learning_rate": 2.7216569767441864e-05,
      "loss": 1.437,
      "mean_token_accuracy": 0.7340640306472779,
      "num_tokens": 53889451.0,
      "step": 750
    },
    {
      "entropy": 0.9999772995710373,
      "epoch": 0.20714091032979012,
      "grad_norm": 0.38552534580230713,
      "learning_rate": 2.757994186046512e-05,
      "loss": 1.4522,
      "mean_token_accuracy": 0.7337017595767975,
      "num_tokens": 54615816.0,
      "step": 760
    },
    {
      "entropy": 0.9884057164192199,
      "epoch": 0.20986644862360315,
      "grad_norm": 0.5651298761367798,
      "learning_rate": 2.7943313953488376e-05,
      "loss": 1.4318,
      "mean_token_accuracy": 0.7341745033860206,
      "num_tokens": 55333318.0,
      "step": 770
    },
    {
      "entropy": 0.9878245323896409,
      "epoch": 0.21259198691741618,
      "grad_norm": 0.4521214962005615,
      "learning_rate": 2.8306686046511626e-05,
      "loss": 1.4445,
      "mean_token_accuracy": 0.7352827414870262,
      "num_tokens": 56063608.0,
      "step": 780
    },
    {
      "entropy": 0.9913873270153999,
      "epoch": 0.2153175252112292,
      "grad_norm": 0.3069559633731842,
      "learning_rate": 2.8670058139534882e-05,
      "loss": 1.4367,
      "mean_token_accuracy": 0.7377872183918953,
      "num_tokens": 56790304.0,
      "step": 790
    },
    {
      "entropy": 0.994021512567997,
      "epoch": 0.21804306350504224,
      "grad_norm": 0.3060591220855713,
      "learning_rate": 2.9033430232558138e-05,
      "loss": 1.4443,
      "mean_token_accuracy": 0.7326282814145089,
      "num_tokens": 57505532.0,
      "step": 800
    },
    {
      "entropy": 0.986484457552433,
      "epoch": 0.22076860179885527,
      "grad_norm": 0.31218984723091125,
      "learning_rate": 2.9396802325581398e-05,
      "loss": 1.4372,
      "mean_token_accuracy": 0.7342789217829704,
      "num_tokens": 58220392.0,
      "step": 810
    },
    {
      "entropy": 0.9901561468839646,
      "epoch": 0.2234941400926683,
      "grad_norm": 0.3456006944179535,
      "learning_rate": 2.9760174418604654e-05,
      "loss": 1.4398,
      "mean_token_accuracy": 0.7340988770127297,
      "num_tokens": 58949816.0,
      "step": 820
    },
    {
      "entropy": 0.9948469802737236,
      "epoch": 0.22621967838648133,
      "grad_norm": 0.32412171363830566,
      "learning_rate": 3.012354651162791e-05,
      "loss": 1.4511,
      "mean_token_accuracy": 0.7328571319580078,
      "num_tokens": 59666066.0,
      "step": 830
    },
    {
      "entropy": 0.9963825002312661,
      "epoch": 0.22894521668029436,
      "grad_norm": 1.0188618898391724,
      "learning_rate": 3.0486918604651166e-05,
      "loss": 1.4426,
      "mean_token_accuracy": 0.730763690173626,
      "num_tokens": 60371355.0,
      "step": 840
    },
    {
      "entropy": 0.9887682050466537,
      "epoch": 0.23167075497410738,
      "grad_norm": 0.3284507095813751,
      "learning_rate": 3.085029069767442e-05,
      "loss": 1.4462,
      "mean_token_accuracy": 0.7367573201656341,
      "num_tokens": 61085148.0,
      "step": 850
    },
    {
      "entropy": 0.9956839889287948,
      "epoch": 0.2343962932679204,
      "grad_norm": 0.25505349040031433,
      "learning_rate": 3.121366279069768e-05,
      "loss": 1.4404,
      "mean_token_accuracy": 0.7327555000782013,
      "num_tokens": 61814554.0,
      "step": 860
    },
    {
      "entropy": 0.9931179091334343,
      "epoch": 0.23712183156173344,
      "grad_norm": 0.3233283460140228,
      "learning_rate": 3.157703488372093e-05,
      "loss": 1.4413,
      "mean_token_accuracy": 0.7317927330732346,
      "num_tokens": 62530557.0,
      "step": 870
    },
    {
      "entropy": 0.9981883198022843,
      "epoch": 0.23984736985554647,
      "grad_norm": 0.32246145606040955,
      "learning_rate": 3.1940406976744184e-05,
      "loss": 1.4506,
      "mean_token_accuracy": 0.7342352598905564,
      "num_tokens": 63245409.0,
      "step": 880
    },
    {
      "entropy": 0.995264945924282,
      "epoch": 0.2425729081493595,
      "grad_norm": 0.3850777745246887,
      "learning_rate": 3.2303779069767444e-05,
      "loss": 1.4388,
      "mean_token_accuracy": 0.7324204593896866,
      "num_tokens": 63960114.0,
      "step": 890
    },
    {
      "entropy": 0.9954063668847084,
      "epoch": 0.24529844644317253,
      "grad_norm": 0.22569897770881653,
      "learning_rate": 3.26671511627907e-05,
      "loss": 1.4422,
      "mean_token_accuracy": 0.7323945328593254,
      "num_tokens": 64684427.0,
      "step": 900
    },
    {
      "entropy": 0.9864580750465393,
      "epoch": 0.24802398473698556,
      "grad_norm": 0.20401442050933838,
      "learning_rate": 3.3030523255813956e-05,
      "loss": 1.4299,
      "mean_token_accuracy": 0.7411849647760391,
      "num_tokens": 65428854.0,
      "step": 910
    },
    {
      "entropy": 0.9876022458076477,
      "epoch": 0.25074952303079856,
      "grad_norm": 0.29853373765945435,
      "learning_rate": 3.339389534883721e-05,
      "loss": 1.4353,
      "mean_token_accuracy": 0.7375764057040215,
      "num_tokens": 66153399.0,
      "step": 920
    },
    {
      "entropy": 0.999330484867096,
      "epoch": 0.2534750613246116,
      "grad_norm": 0.2890157699584961,
      "learning_rate": 3.375726744186047e-05,
      "loss": 1.4528,
      "mean_token_accuracy": 0.7318306133151055,
      "num_tokens": 66862350.0,
      "step": 930
    },
    {
      "entropy": 0.9852076366543769,
      "epoch": 0.2562005996184246,
      "grad_norm": 0.2660386860370636,
      "learning_rate": 3.412063953488372e-05,
      "loss": 1.4284,
      "mean_token_accuracy": 0.7362406551837921,
      "num_tokens": 67585898.0,
      "step": 940
    },
    {
      "entropy": 0.9841386467218399,
      "epoch": 0.2589261379122377,
      "grad_norm": 0.4517176151275635,
      "learning_rate": 3.4484011627906975e-05,
      "loss": 1.4237,
      "mean_token_accuracy": 0.7396635293960572,
      "num_tokens": 68315323.0,
      "step": 950
    },
    {
      "entropy": 0.9917329683899879,
      "epoch": 0.2616516762060507,
      "grad_norm": 0.23541152477264404,
      "learning_rate": 3.4847383720930234e-05,
      "loss": 1.442,
      "mean_token_accuracy": 0.729252065718174,
      "num_tokens": 69017614.0,
      "step": 960
    },
    {
      "entropy": 0.9919871062040329,
      "epoch": 0.26437721449986373,
      "grad_norm": 0.26427310705184937,
      "learning_rate": 3.521075581395349e-05,
      "loss": 1.4444,
      "mean_token_accuracy": 0.7352699220180512,
      "num_tokens": 69740747.0,
      "step": 970
    },
    {
      "entropy": 0.9945264115929604,
      "epoch": 0.26710275279367673,
      "grad_norm": 0.3032136559486389,
      "learning_rate": 3.5574127906976747e-05,
      "loss": 1.4425,
      "mean_token_accuracy": 0.7310560792684555,
      "num_tokens": 70457628.0,
      "step": 980
    },
    {
      "entropy": 0.9931428045034408,
      "epoch": 0.2698282910874898,
      "grad_norm": 0.32616934180259705,
      "learning_rate": 3.59375e-05,
      "loss": 1.4412,
      "mean_token_accuracy": 0.7312234610319137,
      "num_tokens": 71172680.0,
      "step": 990
    },
    {
      "entropy": 0.9885409861803055,
      "epoch": 0.2725538293813028,
      "grad_norm": 0.48906901478767395,
      "learning_rate": 3.630087209302326e-05,
      "loss": 1.4289,
      "mean_token_accuracy": 0.7374913066625595,
      "num_tokens": 71893680.0,
      "step": 1000
    },
    {
      "entropy": 0.9969916045665741,
      "epoch": 0.27527936767511585,
      "grad_norm": 0.24646152555942535,
      "learning_rate": 3.666424418604651e-05,
      "loss": 1.4385,
      "mean_token_accuracy": 0.7325153753161431,
      "num_tokens": 72614213.0,
      "step": 1010
    },
    {
      "entropy": 0.9937289983034134,
      "epoch": 0.27800490596892885,
      "grad_norm": 0.4907693862915039,
      "learning_rate": 3.702761627906977e-05,
      "loss": 1.439,
      "mean_token_accuracy": 0.7257880479097366,
      "num_tokens": 73312163.0,
      "step": 1020
    },
    {
      "entropy": 0.9726315110921859,
      "epoch": 0.2807304442627419,
      "grad_norm": 0.3276204466819763,
      "learning_rate": 3.7390988372093024e-05,
      "loss": 1.4218,
      "mean_token_accuracy": 0.7396723493933678,
      "num_tokens": 74032914.0,
      "step": 1030
    },
    {
      "entropy": 0.9983386799693108,
      "epoch": 0.2834559825565549,
      "grad_norm": 0.29827114939689636,
      "learning_rate": 3.775436046511628e-05,
      "loss": 1.4417,
      "mean_token_accuracy": 0.7349489763379097,
      "num_tokens": 74750343.0,
      "step": 1040
    },
    {
      "entropy": 0.9868063226342201,
      "epoch": 0.28618152085036797,
      "grad_norm": 0.2375505566596985,
      "learning_rate": 3.811773255813954e-05,
      "loss": 1.4362,
      "mean_token_accuracy": 0.7375910192728042,
      "num_tokens": 75472802.0,
      "step": 1050
    },
    {
      "entropy": 0.9938691481947899,
      "epoch": 0.28890705914418097,
      "grad_norm": 0.2215590924024582,
      "learning_rate": 3.848110465116279e-05,
      "loss": 1.4346,
      "mean_token_accuracy": 0.7330907583236694,
      "num_tokens": 76195299.0,
      "step": 1060
    },
    {
      "entropy": 0.9908524438738823,
      "epoch": 0.291632597437994,
      "grad_norm": 0.21515017747879028,
      "learning_rate": 3.884447674418605e-05,
      "loss": 1.4348,
      "mean_token_accuracy": 0.7335933834314347,
      "num_tokens": 76915756.0,
      "step": 1070
    },
    {
      "entropy": 0.9933411166071892,
      "epoch": 0.294358135731807,
      "grad_norm": 0.2014913111925125,
      "learning_rate": 3.92078488372093e-05,
      "loss": 1.4381,
      "mean_token_accuracy": 0.7324068695306778,
      "num_tokens": 77624920.0,
      "step": 1080
    },
    {
      "entropy": 0.9938898369669914,
      "epoch": 0.2970836740256201,
      "grad_norm": 0.41872668266296387,
      "learning_rate": 3.957122093023256e-05,
      "loss": 1.4394,
      "mean_token_accuracy": 0.7363974317908287,
      "num_tokens": 78354545.0,
      "step": 1090
    },
    {
      "entropy": 0.9886097267270089,
      "epoch": 0.2998092123194331,
      "grad_norm": 0.4530754089355469,
      "learning_rate": 3.9934593023255814e-05,
      "loss": 1.432,
      "mean_token_accuracy": 0.740429712831974,
      "num_tokens": 79097138.0,
      "step": 1100
    },
    {
      "entropy": 0.9957090765237808,
      "epoch": 0.30253475061324614,
      "grad_norm": 0.29355210065841675,
      "learning_rate": 4.0297965116279074e-05,
      "loss": 1.4452,
      "mean_token_accuracy": 0.7312169045209884,
      "num_tokens": 79802002.0,
      "step": 1110
    },
    {
      "entropy": 0.9857747182250023,
      "epoch": 0.30526028890705914,
      "grad_norm": 0.2174595147371292,
      "learning_rate": 4.066133720930233e-05,
      "loss": 1.4273,
      "mean_token_accuracy": 0.7381820440292358,
      "num_tokens": 80529799.0,
      "step": 1120
    },
    {
      "entropy": 0.9758066356182098,
      "epoch": 0.3079858272008722,
      "grad_norm": 0.26723766326904297,
      "learning_rate": 4.102470930232558e-05,
      "loss": 1.4239,
      "mean_token_accuracy": 0.7377608880400658,
      "num_tokens": 81251980.0,
      "step": 1130
    },
    {
      "entropy": 0.9935255914926528,
      "epoch": 0.3107113654946852,
      "grad_norm": 0.21104010939598083,
      "learning_rate": 4.138808139534884e-05,
      "loss": 1.4367,
      "mean_token_accuracy": 0.7398475959897042,
      "num_tokens": 81996411.0,
      "step": 1140
    },
    {
      "entropy": 0.9991440311074257,
      "epoch": 0.31343690378849826,
      "grad_norm": 0.28874218463897705,
      "learning_rate": 4.175145348837209e-05,
      "loss": 1.4452,
      "mean_token_accuracy": 0.7331483393907547,
      "num_tokens": 82707313.0,
      "step": 1150
    },
    {
      "entropy": 0.9936499938368797,
      "epoch": 0.31616244208231126,
      "grad_norm": 0.22870078682899475,
      "learning_rate": 4.211482558139535e-05,
      "loss": 1.4373,
      "mean_token_accuracy": 0.7290183573961257,
      "num_tokens": 83410115.0,
      "step": 1160
    },
    {
      "entropy": 0.984348925948143,
      "epoch": 0.31888798037612426,
      "grad_norm": 0.4662875831127167,
      "learning_rate": 4.2478197674418604e-05,
      "loss": 1.4237,
      "mean_token_accuracy": 0.739226421713829,
      "num_tokens": 84141229.0,
      "step": 1170
    },
    {
      "entropy": 0.9941943392157555,
      "epoch": 0.3216135186699373,
      "grad_norm": 0.2791188061237335,
      "learning_rate": 4.2841569767441864e-05,
      "loss": 1.4402,
      "mean_token_accuracy": 0.7304043978452682,
      "num_tokens": 84853257.0,
      "step": 1180
    },
    {
      "entropy": 0.9998081073164939,
      "epoch": 0.3243390569637503,
      "grad_norm": 0.523344874382019,
      "learning_rate": 4.320494186046512e-05,
      "loss": 1.4393,
      "mean_token_accuracy": 0.7313911184668541,
      "num_tokens": 85569444.0,
      "step": 1190
    },
    {
      "entropy": 0.9841462850570679,
      "epoch": 0.3270645952575634,
      "grad_norm": 0.2346402406692505,
      "learning_rate": 4.3568313953488377e-05,
      "loss": 1.4256,
      "mean_token_accuracy": 0.7391437903046608,
      "num_tokens": 86301942.0,
      "step": 1200
    },
    {
      "entropy": 0.9866098940372467,
      "epoch": 0.3297901335513764,
      "grad_norm": 0.28725212812423706,
      "learning_rate": 4.393168604651163e-05,
      "loss": 1.4348,
      "mean_token_accuracy": 0.7339578330516815,
      "num_tokens": 87021466.0,
      "step": 1210
    },
    {
      "entropy": 0.9928266644477844,
      "epoch": 0.33251567184518943,
      "grad_norm": 0.2333146333694458,
      "learning_rate": 4.429505813953488e-05,
      "loss": 1.4294,
      "mean_token_accuracy": 0.7346134960651398,
      "num_tokens": 87750906.0,
      "step": 1220
    },
    {
      "entropy": 0.9935041859745979,
      "epoch": 0.33524121013900243,
      "grad_norm": 0.2144048660993576,
      "learning_rate": 4.465843023255814e-05,
      "loss": 1.4372,
      "mean_token_accuracy": 0.7343584313988686,
      "num_tokens": 88468886.0,
      "step": 1230
    },
    {
      "entropy": 0.9863809257745743,
      "epoch": 0.3379667484328155,
      "grad_norm": 0.19342583417892456,
      "learning_rate": 4.5021802325581395e-05,
      "loss": 1.4381,
      "mean_token_accuracy": 0.7397785291075707,
      "num_tokens": 89192372.0,
      "step": 1240
    },
    {
      "entropy": 0.9911041915416717,
      "epoch": 0.3406922867266285,
      "grad_norm": 0.1899794638156891,
      "learning_rate": 4.5385174418604654e-05,
      "loss": 1.4318,
      "mean_token_accuracy": 0.7350430831313133,
      "num_tokens": 89919594.0,
      "step": 1250
    },
    {
      "entropy": 0.9816926836967468,
      "epoch": 0.34341782502044155,
      "grad_norm": 0.24096062779426575,
      "learning_rate": 4.574854651162791e-05,
      "loss": 1.428,
      "mean_token_accuracy": 0.7362191364169121,
      "num_tokens": 90637594.0,
      "step": 1260
    },
    {
      "entropy": 0.9869430780410766,
      "epoch": 0.34614336331425455,
      "grad_norm": 0.1929672807455063,
      "learning_rate": 4.611191860465117e-05,
      "loss": 1.4318,
      "mean_token_accuracy": 0.7313936963677407,
      "num_tokens": 91349032.0,
      "step": 1270
    },
    {
      "entropy": 0.9814235389232635,
      "epoch": 0.3488689016080676,
      "grad_norm": 0.20565539598464966,
      "learning_rate": 4.6475290697674426e-05,
      "loss": 1.4309,
      "mean_token_accuracy": 0.7382841348648072,
      "num_tokens": 92082961.0,
      "step": 1280
    },
    {
      "entropy": 0.985702796280384,
      "epoch": 0.3515944399018806,
      "grad_norm": 0.21981583535671234,
      "learning_rate": 4.683866279069768e-05,
      "loss": 1.4291,
      "mean_token_accuracy": 0.7318273842334747,
      "num_tokens": 92786151.0,
      "step": 1290
    },
    {
      "entropy": 0.9794762969017029,
      "epoch": 0.35431997819569366,
      "grad_norm": 0.4361553192138672,
      "learning_rate": 4.720203488372093e-05,
      "loss": 1.4314,
      "mean_token_accuracy": 0.7368942067027092,
      "num_tokens": 93503082.0,
      "step": 1300
    },
    {
      "entropy": 0.9880231067538261,
      "epoch": 0.35704551648950666,
      "grad_norm": 0.30570775270462036,
      "learning_rate": 4.7565406976744185e-05,
      "loss": 1.4348,
      "mean_token_accuracy": 0.7372220605611801,
      "num_tokens": 94222464.0,
      "step": 1310
    },
    {
      "entropy": 0.991027458012104,
      "epoch": 0.3597710547833197,
      "grad_norm": 0.34112071990966797,
      "learning_rate": 4.7928779069767444e-05,
      "loss": 1.4406,
      "mean_token_accuracy": 0.7324920043349266,
      "num_tokens": 94925147.0,
      "step": 1320
    },
    {
      "entropy": 0.9888468325138092,
      "epoch": 0.3624965930771327,
      "grad_norm": 0.27501174807548523,
      "learning_rate": 4.82921511627907e-05,
      "loss": 1.4418,
      "mean_token_accuracy": 0.7304517835378647,
      "num_tokens": 95625988.0,
      "step": 1330
    },
    {
      "entropy": 0.9956190213561058,
      "epoch": 0.3652221313709458,
      "grad_norm": 0.4416481852531433,
      "learning_rate": 4.865552325581396e-05,
      "loss": 1.4316,
      "mean_token_accuracy": 0.7305059120059013,
      "num_tokens": 96336107.0,
      "step": 1340
    },
    {
      "entropy": 0.9966611221432686,
      "epoch": 0.3679476696647588,
      "grad_norm": 0.5676621794700623,
      "learning_rate": 4.901889534883721e-05,
      "loss": 1.4485,
      "mean_token_accuracy": 0.7301196932792664,
      "num_tokens": 97043941.0,
      "step": 1350
    },
    {
      "entropy": 0.9922414407134056,
      "epoch": 0.37067320795857184,
      "grad_norm": 0.23170824348926544,
      "learning_rate": 4.938226744186047e-05,
      "loss": 1.4427,
      "mean_token_accuracy": 0.733326929807663,
      "num_tokens": 97755541.0,
      "step": 1360
    },
    {
      "entropy": 0.9969855561852455,
      "epoch": 0.37339874625238484,
      "grad_norm": 0.2700681984424591,
      "learning_rate": 4.974563953488373e-05,
      "loss": 1.4381,
      "mean_token_accuracy": 0.7339515954256057,
      "num_tokens": 98488746.0,
      "step": 1370
    },
    {
      "entropy": 0.9885864198207855,
      "epoch": 0.3761242845461979,
      "grad_norm": 0.24769966304302216,
      "learning_rate": 5.0109011627906975e-05,
      "loss": 1.4445,
      "mean_token_accuracy": 0.7381105780601501,
      "num_tokens": 99213305.0,
      "step": 1380
    },
    {
      "entropy": 0.9947968870401382,
      "epoch": 0.3788498228400109,
      "grad_norm": 0.24572551250457764,
      "learning_rate": 5.047238372093024e-05,
      "loss": 1.4379,
      "mean_token_accuracy": 0.7360276535153389,
      "num_tokens": 99935486.0,
      "step": 1390
    },
    {
      "entropy": 0.9968233689665794,
      "epoch": 0.38157536113382395,
      "grad_norm": 0.2169007509946823,
      "learning_rate": 5.083575581395349e-05,
      "loss": 1.4434,
      "mean_token_accuracy": 0.7335814267396927,
      "num_tokens": 100649239.0,
      "step": 1400
    },
    {
      "entropy": 0.9841761410236358,
      "epoch": 0.38430089942763695,
      "grad_norm": 0.28508707880973816,
      "learning_rate": 5.1199127906976754e-05,
      "loss": 1.426,
      "mean_token_accuracy": 0.730169078707695,
      "num_tokens": 101355919.0,
      "step": 1410
    },
    {
      "entropy": 0.9974832609295845,
      "epoch": 0.38702643772145,
      "grad_norm": 0.21387140452861786,
      "learning_rate": 5.15625e-05,
      "loss": 1.4303,
      "mean_token_accuracy": 0.7316822305321693,
      "num_tokens": 102074622.0,
      "step": 1420
    },
    {
      "entropy": 0.9878896534442901,
      "epoch": 0.389751976015263,
      "grad_norm": 0.21567605435848236,
      "learning_rate": 5.192587209302325e-05,
      "loss": 1.4277,
      "mean_token_accuracy": 0.7342279896140098,
      "num_tokens": 102801841.0,
      "step": 1430
    },
    {
      "entropy": 0.9944760173559188,
      "epoch": 0.39247751430907607,
      "grad_norm": 0.2036101371049881,
      "learning_rate": 5.228924418604651e-05,
      "loss": 1.4236,
      "mean_token_accuracy": 0.7351288288831711,
      "num_tokens": 103529636.0,
      "step": 1440
    },
    {
      "entropy": 0.9866831496357917,
      "epoch": 0.39520305260288907,
      "grad_norm": 0.28441524505615234,
      "learning_rate": 5.2652616279069765e-05,
      "loss": 1.4264,
      "mean_token_accuracy": 0.7344724372029304,
      "num_tokens": 104243246.0,
      "step": 1450
    },
    {
      "entropy": 0.9766260460019112,
      "epoch": 0.3979285908967021,
      "grad_norm": 0.23722435534000397,
      "learning_rate": 5.301598837209303e-05,
      "loss": 1.4239,
      "mean_token_accuracy": 0.7368274286389351,
      "num_tokens": 104948989.0,
      "step": 1460
    },
    {
      "entropy": 0.984954971075058,
      "epoch": 0.40065412919051513,
      "grad_norm": 0.31182342767715454,
      "learning_rate": 5.337936046511628e-05,
      "loss": 1.4304,
      "mean_token_accuracy": 0.734859699010849,
      "num_tokens": 105658858.0,
      "step": 1470
    },
    {
      "entropy": 0.9924835756421089,
      "epoch": 0.40337966748432813,
      "grad_norm": 0.2507476508617401,
      "learning_rate": 5.3742732558139544e-05,
      "loss": 1.436,
      "mean_token_accuracy": 0.7331453621387481,
      "num_tokens": 106382663.0,
      "step": 1480
    },
    {
      "entropy": 0.9932640939950943,
      "epoch": 0.4061052057781412,
      "grad_norm": 0.19360320270061493,
      "learning_rate": 5.410610465116279e-05,
      "loss": 1.4326,
      "mean_token_accuracy": 0.7340785205364228,
      "num_tokens": 107100337.0,
      "step": 1490
    },
    {
      "entropy": 0.9841107860207557,
      "epoch": 0.4088307440719542,
      "grad_norm": 0.18797026574611664,
      "learning_rate": 5.4469476744186056e-05,
      "loss": 1.4328,
      "mean_token_accuracy": 0.7359195277094841,
      "num_tokens": 107815259.0,
      "step": 1500
    },
    {
      "entropy": 0.9935507267713547,
      "epoch": 0.41155628236576725,
      "grad_norm": 0.1644987165927887,
      "learning_rate": 5.48328488372093e-05,
      "loss": 1.4347,
      "mean_token_accuracy": 0.7332268089056015,
      "num_tokens": 108526347.0,
      "step": 1510
    },
    {
      "entropy": 0.9960081830620766,
      "epoch": 0.41428182065958025,
      "grad_norm": 0.15973089635372162,
      "learning_rate": 5.5196220930232555e-05,
      "loss": 1.4446,
      "mean_token_accuracy": 0.7294234141707421,
      "num_tokens": 109227786.0,
      "step": 1520
    },
    {
      "entropy": 0.9813835769891739,
      "epoch": 0.4170073589533933,
      "grad_norm": 0.2454165369272232,
      "learning_rate": 5.555959302325582e-05,
      "loss": 1.4188,
      "mean_token_accuracy": 0.7327255427837371,
      "num_tokens": 109943464.0,
      "step": 1530
    },
    {
      "entropy": 0.9830124020576477,
      "epoch": 0.4197328972472063,
      "grad_norm": 0.300631582736969,
      "learning_rate": 5.592296511627907e-05,
      "loss": 1.429,
      "mean_token_accuracy": 0.7393194779753685,
      "num_tokens": 110673065.0,
      "step": 1540
    },
    {
      "entropy": 0.9851121515035629,
      "epoch": 0.42245843554101936,
      "grad_norm": 0.310230553150177,
      "learning_rate": 5.6286337209302334e-05,
      "loss": 1.4265,
      "mean_token_accuracy": 0.7327947586774826,
      "num_tokens": 111373253.0,
      "step": 1550
    },
    {
      "entropy": 0.9851061567664147,
      "epoch": 0.42518397383483236,
      "grad_norm": 0.2734837830066681,
      "learning_rate": 5.664970930232558e-05,
      "loss": 1.4317,
      "mean_token_accuracy": 0.7363707885146141,
      "num_tokens": 112084685.0,
      "step": 1560
    },
    {
      "entropy": 0.9825420245528221,
      "epoch": 0.4279095121286454,
      "grad_norm": 0.1628209948539734,
      "learning_rate": 5.7013081395348846e-05,
      "loss": 1.4338,
      "mean_token_accuracy": 0.7380750298500061,
      "num_tokens": 112808068.0,
      "step": 1570
    },
    {
      "entropy": 0.9882109865546227,
      "epoch": 0.4306350504224584,
      "grad_norm": 0.19374525547027588,
      "learning_rate": 5.737645348837209e-05,
      "loss": 1.4334,
      "mean_token_accuracy": 0.7342119246721268,
      "num_tokens": 113515057.0,
      "step": 1580
    },
    {
      "entropy": 0.984755776822567,
      "epoch": 0.4333605887162715,
      "grad_norm": 0.2155279815196991,
      "learning_rate": 5.773982558139536e-05,
      "loss": 1.4251,
      "mean_token_accuracy": 0.7343093782663346,
      "num_tokens": 114232069.0,
      "step": 1590
    },
    {
      "entropy": 0.9929274395108223,
      "epoch": 0.4360861270100845,
      "grad_norm": 0.4078405797481537,
      "learning_rate": 5.8103197674418605e-05,
      "loss": 1.4327,
      "mean_token_accuracy": 0.7307331010699272,
      "num_tokens": 114938908.0,
      "step": 1600
    },
    {
      "entropy": 0.9821958124637604,
      "epoch": 0.43881166530389754,
      "grad_norm": 0.2290513664484024,
      "learning_rate": 5.846656976744186e-05,
      "loss": 1.4327,
      "mean_token_accuracy": 0.7371584936976433,
      "num_tokens": 115652920.0,
      "step": 1610
    },
    {
      "entropy": 0.9950265571475029,
      "epoch": 0.44153720359771054,
      "grad_norm": 0.2519257068634033,
      "learning_rate": 5.8829941860465124e-05,
      "loss": 1.435,
      "mean_token_accuracy": 0.7324141755700111,
      "num_tokens": 116363475.0,
      "step": 1620
    },
    {
      "entropy": 0.9973052412271499,
      "epoch": 0.4442627418915236,
      "grad_norm": 0.2373785674571991,
      "learning_rate": 5.919331395348837e-05,
      "loss": 1.4325,
      "mean_token_accuracy": 0.7327213540673256,
      "num_tokens": 117091157.0,
      "step": 1630
    },
    {
      "entropy": 0.9813062354922295,
      "epoch": 0.4469882801853366,
      "grad_norm": 0.2199847549200058,
      "learning_rate": 5.9556686046511636e-05,
      "loss": 1.4201,
      "mean_token_accuracy": 0.7371037125587463,
      "num_tokens": 117809062.0,
      "step": 1640
    },
    {
      "entropy": 0.9898816034197807,
      "epoch": 0.44971381847914965,
      "grad_norm": 0.31303584575653076,
      "learning_rate": 5.992005813953488e-05,
      "loss": 1.4323,
      "mean_token_accuracy": 0.7392094865441322,
      "num_tokens": 118544540.0,
      "step": 1650
    },
    {
      "entropy": 0.9923418626189232,
      "epoch": 0.45243935677296265,
      "grad_norm": 0.21581219136714935,
      "learning_rate": 6.028343023255815e-05,
      "loss": 1.4337,
      "mean_token_accuracy": 0.7334046080708504,
      "num_tokens": 119258189.0,
      "step": 1660
    },
    {
      "entropy": 0.9849744439125061,
      "epoch": 0.4551648950667757,
      "grad_norm": 0.2109210193157196,
      "learning_rate": 6.0646802325581395e-05,
      "loss": 1.4271,
      "mean_token_accuracy": 0.7396485283970833,
      "num_tokens": 119982024.0,
      "step": 1670
    },
    {
      "entropy": 0.9821529731154441,
      "epoch": 0.4578904333605887,
      "grad_norm": 0.22036734223365784,
      "learning_rate": 6.101017441860465e-05,
      "loss": 1.4222,
      "mean_token_accuracy": 0.7384767159819603,
      "num_tokens": 120707990.0,
      "step": 1680
    },
    {
      "entropy": 0.9901896312832832,
      "epoch": 0.46061597165440177,
      "grad_norm": 0.19113869965076447,
      "learning_rate": 6.137354651162791e-05,
      "loss": 1.4303,
      "mean_token_accuracy": 0.731666874885559,
      "num_tokens": 121427857.0,
      "step": 1690
    },
    {
      "entropy": 0.9816726014018059,
      "epoch": 0.46334150994821477,
      "grad_norm": 0.20121638476848602,
      "learning_rate": 6.173691860465116e-05,
      "loss": 1.4286,
      "mean_token_accuracy": 0.7375072449445724,
      "num_tokens": 122141315.0,
      "step": 1700
    },
    {
      "entropy": 0.9800962820649147,
      "epoch": 0.4660670482420278,
      "grad_norm": 0.27943041920661926,
      "learning_rate": 6.210029069767442e-05,
      "loss": 1.4289,
      "mean_token_accuracy": 0.735170379281044,
      "num_tokens": 122845583.0,
      "step": 1710
    },
    {
      "entropy": 0.9852249696850777,
      "epoch": 0.4687925865358408,
      "grad_norm": 0.25668081641197205,
      "learning_rate": 6.246366279069768e-05,
      "loss": 1.4333,
      "mean_token_accuracy": 0.7355706915259361,
      "num_tokens": 123566283.0,
      "step": 1720
    },
    {
      "entropy": 0.9883533254265785,
      "epoch": 0.4715181248296539,
      "grad_norm": 0.18696428835391998,
      "learning_rate": 6.282703488372094e-05,
      "loss": 1.4323,
      "mean_token_accuracy": 0.7344200700521469,
      "num_tokens": 124278516.0,
      "step": 1730
    },
    {
      "entropy": 0.9870807081460953,
      "epoch": 0.4742436631234669,
      "grad_norm": 0.19319681823253632,
      "learning_rate": 6.319040697674419e-05,
      "loss": 1.4277,
      "mean_token_accuracy": 0.7365616247057915,
      "num_tokens": 125010952.0,
      "step": 1740
    },
    {
      "entropy": 0.9897187322378158,
      "epoch": 0.4769692014172799,
      "grad_norm": 0.3825794756412506,
      "learning_rate": 6.355377906976744e-05,
      "loss": 1.4341,
      "mean_token_accuracy": 0.7344249919056892,
      "num_tokens": 125735883.0,
      "step": 1750
    },
    {
      "entropy": 0.9973450332880021,
      "epoch": 0.47969473971109294,
      "grad_norm": 0.20322610437870026,
      "learning_rate": 6.39171511627907e-05,
      "loss": 1.4361,
      "mean_token_accuracy": 0.7336152747273446,
      "num_tokens": 126457176.0,
      "step": 1760
    },
    {
      "entropy": 0.9879305228590965,
      "epoch": 0.48242027800490594,
      "grad_norm": 0.16834957897663116,
      "learning_rate": 6.428052325581395e-05,
      "loss": 1.4297,
      "mean_token_accuracy": 0.7342620983719825,
      "num_tokens": 127178073.0,
      "step": 1770
    },
    {
      "entropy": 0.9840679958462715,
      "epoch": 0.485145816298719,
      "grad_norm": 0.38834550976753235,
      "learning_rate": 6.464389534883721e-05,
      "loss": 1.4318,
      "mean_token_accuracy": 0.7370759963989257,
      "num_tokens": 127898699.0,
      "step": 1780
    },
    {
      "entropy": 0.9803714036941529,
      "epoch": 0.487871354592532,
      "grad_norm": 0.2102237194776535,
      "learning_rate": 6.500726744186047e-05,
      "loss": 1.4293,
      "mean_token_accuracy": 0.735047759115696,
      "num_tokens": 128612251.0,
      "step": 1790
    },
    {
      "entropy": 0.9915691018104553,
      "epoch": 0.49059689288634506,
      "grad_norm": 0.16007673740386963,
      "learning_rate": 6.537063953488373e-05,
      "loss": 1.4371,
      "mean_token_accuracy": 0.7352234482765198,
      "num_tokens": 129329003.0,
      "step": 1800
    },
    {
      "entropy": 0.9829947292804718,
      "epoch": 0.49332243118015806,
      "grad_norm": 0.1799066960811615,
      "learning_rate": 6.573401162790698e-05,
      "loss": 1.4235,
      "mean_token_accuracy": 0.7371339723467827,
      "num_tokens": 130049818.0,
      "step": 1810
    },
    {
      "entropy": 0.9902053967118263,
      "epoch": 0.4960479694739711,
      "grad_norm": 0.18206138908863068,
      "learning_rate": 6.609738372093023e-05,
      "loss": 1.4363,
      "mean_token_accuracy": 0.7356316760182381,
      "num_tokens": 130770612.0,
      "step": 1820
    },
    {
      "entropy": 0.981654267013073,
      "epoch": 0.4987735077677841,
      "grad_norm": 0.288813978433609,
      "learning_rate": 6.64607558139535e-05,
      "loss": 1.429,
      "mean_token_accuracy": 0.7381418347358704,
      "num_tokens": 131489428.0,
      "step": 1830
    },
    {
      "entropy": 0.9810149937868118,
      "epoch": 0.5014990460615971,
      "grad_norm": 0.1831810176372528,
      "learning_rate": 6.682412790697675e-05,
      "loss": 1.422,
      "mean_token_accuracy": 0.7390474632382393,
      "num_tokens": 132211947.0,
      "step": 1840
    },
    {
      "entropy": 0.9969974964857101,
      "epoch": 0.5042245843554102,
      "grad_norm": 0.24637824296951294,
      "learning_rate": 6.71875e-05,
      "loss": 1.4378,
      "mean_token_accuracy": 0.7319372579455375,
      "num_tokens": 132922541.0,
      "step": 1850
    },
    {
      "entropy": 0.9870500966906548,
      "epoch": 0.5069501226492232,
      "grad_norm": 0.2491854578256607,
      "learning_rate": 6.755087209302326e-05,
      "loss": 1.4254,
      "mean_token_accuracy": 0.7341958448290825,
      "num_tokens": 133643346.0,
      "step": 1860
    },
    {
      "entropy": 0.9897508502006531,
      "epoch": 0.5096756609430363,
      "grad_norm": 0.18362078070640564,
      "learning_rate": 6.791424418604652e-05,
      "loss": 1.4282,
      "mean_token_accuracy": 0.7369514137506485,
      "num_tokens": 134373792.0,
      "step": 1870
    },
    {
      "entropy": 0.9835589408874512,
      "epoch": 0.5124011992368492,
      "grad_norm": 0.21839305758476257,
      "learning_rate": 6.827761627906977e-05,
      "loss": 1.4232,
      "mean_token_accuracy": 0.738610865175724,
      "num_tokens": 135095994.0,
      "step": 1880
    },
    {
      "entropy": 0.9815672129392624,
      "epoch": 0.5151267375306623,
      "grad_norm": 0.21944372355937958,
      "learning_rate": 6.864098837209302e-05,
      "loss": 1.424,
      "mean_token_accuracy": 0.7349463656544686,
      "num_tokens": 135804246.0,
      "step": 1890
    },
    {
      "entropy": 0.9870554804801941,
      "epoch": 0.5178522758244753,
      "grad_norm": 0.18906302750110626,
      "learning_rate": 6.900436046511628e-05,
      "loss": 1.4238,
      "mean_token_accuracy": 0.7398572236299514,
      "num_tokens": 136536781.0,
      "step": 1900
    },
    {
      "entropy": 0.9853287354111672,
      "epoch": 0.5205778141182884,
      "grad_norm": 0.18919885158538818,
      "learning_rate": 6.936773255813954e-05,
      "loss": 1.428,
      "mean_token_accuracy": 0.7393600136041641,
      "num_tokens": 137266151.0,
      "step": 1910
    },
    {
      "entropy": 0.9780439928174018,
      "epoch": 0.5233033524121014,
      "grad_norm": 0.1842227280139923,
      "learning_rate": 6.973110465116279e-05,
      "loss": 1.4242,
      "mean_token_accuracy": 0.7379230797290802,
      "num_tokens": 137975510.0,
      "step": 1920
    },
    {
      "entropy": 0.9827481582760811,
      "epoch": 0.5260288907059144,
      "grad_norm": 0.2542431056499481,
      "learning_rate": 7.009447674418605e-05,
      "loss": 1.4187,
      "mean_token_accuracy": 0.7364926964044571,
      "num_tokens": 138699820.0,
      "step": 1930
    },
    {
      "entropy": 0.983500225841999,
      "epoch": 0.5287544289997275,
      "grad_norm": 0.16537941992282867,
      "learning_rate": 7.045784883720931e-05,
      "loss": 1.4253,
      "mean_token_accuracy": 0.7361272245645523,
      "num_tokens": 139421211.0,
      "step": 1940
    },
    {
      "entropy": 0.9835821121931076,
      "epoch": 0.5314799672935405,
      "grad_norm": 0.1662430763244629,
      "learning_rate": 7.082122093023256e-05,
      "loss": 1.4259,
      "mean_token_accuracy": 0.7393700927495956,
      "num_tokens": 140143156.0,
      "step": 1950
    },
    {
      "entropy": 0.9743385061621666,
      "epoch": 0.5342055055873535,
      "grad_norm": 0.1770647168159485,
      "learning_rate": 7.118459302325582e-05,
      "loss": 1.4193,
      "mean_token_accuracy": 0.7395164668560028,
      "num_tokens": 140864472.0,
      "step": 1960
    },
    {
      "entropy": 0.9933014005422592,
      "epoch": 0.5369310438811665,
      "grad_norm": 0.17968490719795227,
      "learning_rate": 7.154796511627907e-05,
      "loss": 1.4407,
      "mean_token_accuracy": 0.7328062981367112,
      "num_tokens": 141571085.0,
      "step": 1970
    },
    {
      "entropy": 0.9851297155022621,
      "epoch": 0.5396565821749796,
      "grad_norm": 0.17354978621006012,
      "learning_rate": 7.191133720930233e-05,
      "loss": 1.4192,
      "mean_token_accuracy": 0.7395014494657517,
      "num_tokens": 142303250.0,
      "step": 1980
    },
    {
      "entropy": 0.9806404992938041,
      "epoch": 0.5423821204687926,
      "grad_norm": 0.15029194951057434,
      "learning_rate": 7.227470930232558e-05,
      "loss": 1.4173,
      "mean_token_accuracy": 0.7421912506222725,
      "num_tokens": 143047352.0,
      "step": 1990
    },
    {
      "entropy": 0.9813976526260376,
      "epoch": 0.5451076587626056,
      "grad_norm": 0.1776154786348343,
      "learning_rate": 7.263808139534884e-05,
      "loss": 1.4225,
      "mean_token_accuracy": 0.7393262058496475,
      "num_tokens": 143760780.0,
      "step": 2000
    },
    {
      "entropy": 0.9802085116505623,
      "epoch": 0.5478331970564186,
      "grad_norm": 0.14664503931999207,
      "learning_rate": 7.30014534883721e-05,
      "loss": 1.4155,
      "mean_token_accuracy": 0.7426095440983772,
      "num_tokens": 144497473.0,
      "step": 2010
    },
    {
      "entropy": 0.9769001394510269,
      "epoch": 0.5505587353502317,
      "grad_norm": 0.15867289900779724,
      "learning_rate": 7.336482558139536e-05,
      "loss": 1.4173,
      "mean_token_accuracy": 0.7400016963481904,
      "num_tokens": 145230194.0,
      "step": 2020
    },
    {
      "entropy": 0.9801029890775681,
      "epoch": 0.5532842736440446,
      "grad_norm": 0.2603178322315216,
      "learning_rate": 7.37281976744186e-05,
      "loss": 1.4156,
      "mean_token_accuracy": 0.7437982723116875,
      "num_tokens": 145967549.0,
      "step": 2030
    },
    {
      "entropy": 0.9902677729725837,
      "epoch": 0.5560098119378577,
      "grad_norm": 0.17806479334831238,
      "learning_rate": 7.409156976744186e-05,
      "loss": 1.4317,
      "mean_token_accuracy": 0.734989982843399,
      "num_tokens": 146689707.0,
      "step": 2040
    },
    {
      "entropy": 0.9883522406220436,
      "epoch": 0.5587353502316708,
      "grad_norm": 0.19524773955345154,
      "learning_rate": 7.445494186046512e-05,
      "loss": 1.4274,
      "mean_token_accuracy": 0.7380678072571755,
      "num_tokens": 147399527.0,
      "step": 2050
    },
    {
      "entropy": 0.9826523959636688,
      "epoch": 0.5614608885254838,
      "grad_norm": 0.18622492253780365,
      "learning_rate": 7.481831395348837e-05,
      "loss": 1.4259,
      "mean_token_accuracy": 0.739601357281208,
      "num_tokens": 148129697.0,
      "step": 2060
    },
    {
      "entropy": 0.9859274953603745,
      "epoch": 0.5641864268192968,
      "grad_norm": 0.23375269770622253,
      "learning_rate": 7.518168604651163e-05,
      "loss": 1.4272,
      "mean_token_accuracy": 0.739434064924717,
      "num_tokens": 148858960.0,
      "step": 2070
    },
    {
      "entropy": 0.9859586983919144,
      "epoch": 0.5669119651131098,
      "grad_norm": 0.16675522923469543,
      "learning_rate": 7.554505813953489e-05,
      "loss": 1.4265,
      "mean_token_accuracy": 0.7352262407541275,
      "num_tokens": 149585854.0,
      "step": 2080
    },
    {
      "entropy": 0.9837883651256562,
      "epoch": 0.5696375034069229,
      "grad_norm": 0.2244209200143814,
      "learning_rate": 7.590843023255815e-05,
      "loss": 1.421,
      "mean_token_accuracy": 0.7335263103246689,
      "num_tokens": 150305785.0,
      "step": 2090
    },
    {
      "entropy": 0.9847618371248246,
      "epoch": 0.5723630417007359,
      "grad_norm": 0.2683162987232208,
      "learning_rate": 7.62718023255814e-05,
      "loss": 1.4176,
      "mean_token_accuracy": 0.7359615460038185,
      "num_tokens": 151023546.0,
      "step": 2100
    },
    {
      "entropy": 0.9798925578594208,
      "epoch": 0.5750885799945489,
      "grad_norm": 0.18732650578022003,
      "learning_rate": 7.663517441860465e-05,
      "loss": 1.4222,
      "mean_token_accuracy": 0.7367407724261283,
      "num_tokens": 151738755.0,
      "step": 2110
    },
    {
      "entropy": 0.9841523870825768,
      "epoch": 0.5778141182883619,
      "grad_norm": 0.17274878919124603,
      "learning_rate": 7.699854651162791e-05,
      "loss": 1.4206,
      "mean_token_accuracy": 0.736824955046177,
      "num_tokens": 152459628.0,
      "step": 2120
    },
    {
      "entropy": 0.9872769743204117,
      "epoch": 0.580539656582175,
      "grad_norm": 0.22543658316135406,
      "learning_rate": 7.736191860465116e-05,
      "loss": 1.4305,
      "mean_token_accuracy": 0.7352962151169777,
      "num_tokens": 153179643.0,
      "step": 2130
    },
    {
      "entropy": 0.9773565545678139,
      "epoch": 0.583265194875988,
      "grad_norm": 0.1713961362838745,
      "learning_rate": 7.772529069767442e-05,
      "loss": 1.4141,
      "mean_token_accuracy": 0.7409474954009057,
      "num_tokens": 153898819.0,
      "step": 2140
    },
    {
      "entropy": 0.9789120152592659,
      "epoch": 0.585990733169801,
      "grad_norm": 0.5829377174377441,
      "learning_rate": 7.808866279069768e-05,
      "loss": 1.4283,
      "mean_token_accuracy": 0.7368653923273086,
      "num_tokens": 154607747.0,
      "step": 2150
    },
    {
      "entropy": 0.9918661847710609,
      "epoch": 0.588716271463614,
      "grad_norm": 0.18343167006969452,
      "learning_rate": 7.845203488372094e-05,
      "loss": 1.4353,
      "mean_token_accuracy": 0.733272735774517,
      "num_tokens": 155331252.0,
      "step": 2160
    },
    {
      "entropy": 0.9797570928931236,
      "epoch": 0.5914418097574271,
      "grad_norm": 0.16955235600471497,
      "learning_rate": 7.881540697674419e-05,
      "loss": 1.4191,
      "mean_token_accuracy": 0.7337436139583587,
      "num_tokens": 156039176.0,
      "step": 2170
    },
    {
      "entropy": 0.9858640789985657,
      "epoch": 0.5941673480512402,
      "grad_norm": 0.14670498669147491,
      "learning_rate": 7.917877906976744e-05,
      "loss": 1.4148,
      "mean_token_accuracy": 0.7360583290457725,
      "num_tokens": 156768336.0,
      "step": 2180
    },
    {
      "entropy": 0.9773233219981193,
      "epoch": 0.5968928863450531,
      "grad_norm": 0.2751343250274658,
      "learning_rate": 7.95421511627907e-05,
      "loss": 1.4145,
      "mean_token_accuracy": 0.7375390321016312,
      "num_tokens": 157477191.0,
      "step": 2190
    },
    {
      "entropy": 0.9836308777332305,
      "epoch": 0.5996184246388662,
      "grad_norm": 0.16778914630413055,
      "learning_rate": 7.990552325581395e-05,
      "loss": 1.4319,
      "mean_token_accuracy": 0.7302121177315712,
      "num_tokens": 158171395.0,
      "step": 2200
    },
    {
      "entropy": 0.9806986883282661,
      "epoch": 0.6023439629326792,
      "grad_norm": 0.21272782981395721,
      "learning_rate": 8.026889534883721e-05,
      "loss": 1.4204,
      "mean_token_accuracy": 0.7391786813735962,
      "num_tokens": 158900598.0,
      "step": 2210
    },
    {
      "entropy": 0.9803844973444938,
      "epoch": 0.6050695012264923,
      "grad_norm": 0.21859610080718994,
      "learning_rate": 8.063226744186047e-05,
      "loss": 1.4211,
      "mean_token_accuracy": 0.7387867569923401,
      "num_tokens": 159624174.0,
      "step": 2220
    },
    {
      "entropy": 0.9872574612498284,
      "epoch": 0.6077950395203052,
      "grad_norm": 0.24079576134681702,
      "learning_rate": 8.099563953488373e-05,
      "loss": 1.4164,
      "mean_token_accuracy": 0.7402409985661507,
      "num_tokens": 160352725.0,
      "step": 2230
    },
    {
      "entropy": 0.9936911195516587,
      "epoch": 0.6105205778141183,
      "grad_norm": 0.2104625403881073,
      "learning_rate": 8.135901162790698e-05,
      "loss": 1.4254,
      "mean_token_accuracy": 0.7338799551129341,
      "num_tokens": 161070338.0,
      "step": 2240
    },
    {
      "entropy": 0.9814269945025444,
      "epoch": 0.6132461161079313,
      "grad_norm": 0.12204383313655853,
      "learning_rate": 8.172238372093024e-05,
      "loss": 1.422,
      "mean_token_accuracy": 0.7419388547539711,
      "num_tokens": 161808253.0,
      "step": 2250
    },
    {
      "entropy": 0.9859389901161194,
      "epoch": 0.6159716544017444,
      "grad_norm": 0.2447647750377655,
      "learning_rate": 8.20857558139535e-05,
      "loss": 1.4288,
      "mean_token_accuracy": 0.7386970460414887,
      "num_tokens": 162526434.0,
      "step": 2260
    },
    {
      "entropy": 0.9854311719536781,
      "epoch": 0.6186971926955573,
      "grad_norm": 0.22936969995498657,
      "learning_rate": 8.244912790697675e-05,
      "loss": 1.4248,
      "mean_token_accuracy": 0.73409523665905,
      "num_tokens": 163247416.0,
      "step": 2270
    },
    {
      "entropy": 0.9888584405183792,
      "epoch": 0.6214227309893704,
      "grad_norm": 0.1381879299879074,
      "learning_rate": 8.28125e-05,
      "loss": 1.4374,
      "mean_token_accuracy": 0.7325819313526154,
      "num_tokens": 163951723.0,
      "step": 2280
    },
    {
      "entropy": 0.9800671771168709,
      "epoch": 0.6241482692831835,
      "grad_norm": 0.137205109000206,
      "learning_rate": 8.317587209302326e-05,
      "loss": 1.4241,
      "mean_token_accuracy": 0.7337846279144287,
      "num_tokens": 164650972.0,
      "step": 2290
    },
    {
      "entropy": 0.9876679092645645,
      "epoch": 0.6268738075769965,
      "grad_norm": 0.1826333999633789,
      "learning_rate": 8.353924418604652e-05,
      "loss": 1.4297,
      "mean_token_accuracy": 0.7341669753193856,
      "num_tokens": 165361994.0,
      "step": 2300
    },
    {
      "entropy": 0.9780210003256797,
      "epoch": 0.6295993458708095,
      "grad_norm": 0.24412642419338226,
      "learning_rate": 8.390261627906977e-05,
      "loss": 1.4136,
      "mean_token_accuracy": 0.7400619566440583,
      "num_tokens": 166092061.0,
      "step": 2310
    },
    {
      "entropy": 0.9783180683851243,
      "epoch": 0.6323248841646225,
      "grad_norm": 0.26498180627822876,
      "learning_rate": 8.426598837209303e-05,
      "loss": 1.4203,
      "mean_token_accuracy": 0.7401160180568696,
      "num_tokens": 166813871.0,
      "step": 2320
    },
    {
      "entropy": 0.9865237712860108,
      "epoch": 0.6350504224584356,
      "grad_norm": 0.3049558103084564,
      "learning_rate": 8.462936046511628e-05,
      "loss": 1.4243,
      "mean_token_accuracy": 0.7345960304141045,
      "num_tokens": 167527305.0,
      "step": 2330
    },
    {
      "entropy": 0.9801740884780884,
      "epoch": 0.6377759607522485,
      "grad_norm": 0.33257248997688293,
      "learning_rate": 8.499273255813954e-05,
      "loss": 1.4164,
      "mean_token_accuracy": 0.7419341236352921,
      "num_tokens": 168256549.0,
      "step": 2340
    },
    {
      "entropy": 0.9854059845209122,
      "epoch": 0.6405014990460616,
      "grad_norm": 0.15806418657302856,
      "learning_rate": 8.535610465116279e-05,
      "loss": 1.4275,
      "mean_token_accuracy": 0.7322300463914871,
      "num_tokens": 168957455.0,
      "step": 2350
    },
    {
      "entropy": 0.9823917403817177,
      "epoch": 0.6432270373398746,
      "grad_norm": 0.21669892966747284,
      "learning_rate": 8.571947674418605e-05,
      "loss": 1.419,
      "mean_token_accuracy": 0.7398874923586846,
      "num_tokens": 169689713.0,
      "step": 2360
    },
    {
      "entropy": 0.9831691041588784,
      "epoch": 0.6459525756336877,
      "grad_norm": 0.3539750874042511,
      "learning_rate": 8.608284883720931e-05,
      "loss": 1.4203,
      "mean_token_accuracy": 0.7360575824975968,
      "num_tokens": 170413913.0,
      "step": 2370
    },
    {
      "entropy": 0.9930136293172837,
      "epoch": 0.6486781139275006,
      "grad_norm": 0.6360985040664673,
      "learning_rate": 8.644622093023256e-05,
      "loss": 1.4363,
      "mean_token_accuracy": 0.7349853679537773,
      "num_tokens": 171125900.0,
      "step": 2380
    },
    {
      "entropy": 0.980540269613266,
      "epoch": 0.6514036522213137,
      "grad_norm": 0.3049725294113159,
      "learning_rate": 8.680959302325582e-05,
      "loss": 1.4163,
      "mean_token_accuracy": 0.7372217386960983,
      "num_tokens": 171847267.0,
      "step": 2390
    },
    {
      "entropy": 0.9860322341322899,
      "epoch": 0.6541291905151267,
      "grad_norm": 0.21088938415050507,
      "learning_rate": 8.717296511627907e-05,
      "loss": 1.4273,
      "mean_token_accuracy": 0.7365284189581871,
      "num_tokens": 172575435.0,
      "step": 2400
    },
    {
      "entropy": 0.9814476162195206,
      "epoch": 0.6568547288089398,
      "grad_norm": 0.17495831847190857,
      "learning_rate": 8.753633720930233e-05,
      "loss": 1.4252,
      "mean_token_accuracy": 0.7373342961072922,
      "num_tokens": 173287782.0,
      "step": 2410
    },
    {
      "entropy": 0.9866706624627113,
      "epoch": 0.6595802671027527,
      "grad_norm": 0.15505193173885345,
      "learning_rate": 8.789970930232558e-05,
      "loss": 1.4261,
      "mean_token_accuracy": 0.7357053935527802,
      "num_tokens": 174003727.0,
      "step": 2420
    },
    {
      "entropy": 0.9791345074772835,
      "epoch": 0.6623058053965658,
      "grad_norm": 0.13619117438793182,
      "learning_rate": 8.826308139534884e-05,
      "loss": 1.4148,
      "mean_token_accuracy": 0.7366743147373199,
      "num_tokens": 174706233.0,
      "step": 2430
    },
    {
      "entropy": 0.9816969931125641,
      "epoch": 0.6650313436903789,
      "grad_norm": 0.23116159439086914,
      "learning_rate": 8.86264534883721e-05,
      "loss": 1.4232,
      "mean_token_accuracy": 0.7392698466777802,
      "num_tokens": 175429248.0,
      "step": 2440
    },
    {
      "entropy": 0.9880701199173927,
      "epoch": 0.6677568819841919,
      "grad_norm": 0.14319849014282227,
      "learning_rate": 8.898982558139536e-05,
      "loss": 1.4379,
      "mean_token_accuracy": 0.7343179911375046,
      "num_tokens": 176139796.0,
      "step": 2450
    },
    {
      "entropy": 0.9857257276773452,
      "epoch": 0.6704824202780049,
      "grad_norm": 0.2317603975534439,
      "learning_rate": 8.93531976744186e-05,
      "loss": 1.4275,
      "mean_token_accuracy": 0.7344384148716927,
      "num_tokens": 176850673.0,
      "step": 2460
    },
    {
      "entropy": 0.9848071858286858,
      "epoch": 0.6732079585718179,
      "grad_norm": 0.17459312081336975,
      "learning_rate": 8.971656976744187e-05,
      "loss": 1.4287,
      "mean_token_accuracy": 0.7325173929333687,
      "num_tokens": 177563599.0,
      "step": 2470
    },
    {
      "entropy": 0.9801058009266853,
      "epoch": 0.675933496865631,
      "grad_norm": 0.16312763094902039,
      "learning_rate": 9.007994186046512e-05,
      "loss": 1.4318,
      "mean_token_accuracy": 0.7353872433304787,
      "num_tokens": 178263299.0,
      "step": 2480
    },
    {
      "entropy": 0.9788572788238525,
      "epoch": 0.678659035159444,
      "grad_norm": 0.19579234719276428,
      "learning_rate": 9.044331395348837e-05,
      "loss": 1.4239,
      "mean_token_accuracy": 0.7329365327954293,
      "num_tokens": 178971882.0,
      "step": 2490
    },
    {
      "entropy": 0.9752021312713623,
      "epoch": 0.681384573453257,
      "grad_norm": 0.15345396101474762,
      "learning_rate": 9.080668604651163e-05,
      "loss": 1.4224,
      "mean_token_accuracy": 0.7376372799277305,
      "num_tokens": 179678144.0,
      "step": 2500
    },
    {
      "entropy": 0.9790006041526794,
      "epoch": 0.68411011174707,
      "grad_norm": 0.13730672001838684,
      "learning_rate": 9.117005813953489e-05,
      "loss": 1.4215,
      "mean_token_accuracy": 0.7365571096539497,
      "num_tokens": 180396043.0,
      "step": 2510
    },
    {
      "entropy": 0.9847607880830764,
      "epoch": 0.6868356500408831,
      "grad_norm": 0.17317773401737213,
      "learning_rate": 9.153343023255815e-05,
      "loss": 1.4241,
      "mean_token_accuracy": 0.7385125055909156,
      "num_tokens": 181124232.0,
      "step": 2520
    },
    {
      "entropy": 0.9774447113275528,
      "epoch": 0.6895611883346962,
      "grad_norm": 0.21009580790996552,
      "learning_rate": 9.18968023255814e-05,
      "loss": 1.4204,
      "mean_token_accuracy": 0.737717455625534,
      "num_tokens": 181837654.0,
      "step": 2530
    },
    {
      "entropy": 0.984051650762558,
      "epoch": 0.6922867266285091,
      "grad_norm": 0.12575000524520874,
      "learning_rate": 9.226017441860466e-05,
      "loss": 1.4247,
      "mean_token_accuracy": 0.7335541769862175,
      "num_tokens": 182539820.0,
      "step": 2540
    },
    {
      "entropy": 0.9811558112502098,
      "epoch": 0.6950122649223222,
      "grad_norm": 0.1522400677204132,
      "learning_rate": 9.262354651162791e-05,
      "loss": 1.4148,
      "mean_token_accuracy": 0.7406320959329605,
      "num_tokens": 183264633.0,
      "step": 2550
    },
    {
      "entropy": 0.9774328708648682,
      "epoch": 0.6977378032161352,
      "grad_norm": 0.13843339681625366,
      "learning_rate": 9.298691860465116e-05,
      "loss": 1.4128,
      "mean_token_accuracy": 0.7441829890012741,
      "num_tokens": 184000840.0,
      "step": 2560
    },
    {
      "entropy": 0.9893192529678345,
      "epoch": 0.7004633415099483,
      "grad_norm": 0.15118694305419922,
      "learning_rate": 9.335029069767442e-05,
      "loss": 1.4272,
      "mean_token_accuracy": 0.7330075368285179,
      "num_tokens": 184709055.0,
      "step": 2570
    },
    {
      "entropy": 0.982662744820118,
      "epoch": 0.7031888798037612,
      "grad_norm": 0.1605202555656433,
      "learning_rate": 9.371366279069768e-05,
      "loss": 1.4238,
      "mean_token_accuracy": 0.7381146803498269,
      "num_tokens": 185426684.0,
      "step": 2580
    },
    {
      "entropy": 0.9822832211852074,
      "epoch": 0.7059144180975743,
      "grad_norm": 0.1687488704919815,
      "learning_rate": 9.407703488372094e-05,
      "loss": 1.4203,
      "mean_token_accuracy": 0.7379755496978759,
      "num_tokens": 186154172.0,
      "step": 2590
    },
    {
      "entropy": 0.9851330190896987,
      "epoch": 0.7086399563913873,
      "grad_norm": 0.13439348340034485,
      "learning_rate": 9.444040697674419e-05,
      "loss": 1.422,
      "mean_token_accuracy": 0.7387931779026985,
      "num_tokens": 186872408.0,
      "step": 2600
    },
    {
      "entropy": 0.9774848327040673,
      "epoch": 0.7113654946852003,
      "grad_norm": 0.18168923258781433,
      "learning_rate": 9.480377906976745e-05,
      "loss": 1.4232,
      "mean_token_accuracy": 0.7363077029585838,
      "num_tokens": 187586232.0,
      "step": 2610
    },
    {
      "entropy": 0.9839216858148575,
      "epoch": 0.7140910329790133,
      "grad_norm": 0.20491938292980194,
      "learning_rate": 9.51671511627907e-05,
      "loss": 1.4205,
      "mean_token_accuracy": 0.7323327973484993,
      "num_tokens": 188296851.0,
      "step": 2620
    },
    {
      "entropy": 0.9769590377807618,
      "epoch": 0.7168165712728264,
      "grad_norm": 0.1636561155319214,
      "learning_rate": 9.553052325581395e-05,
      "loss": 1.4122,
      "mean_token_accuracy": 0.7377440124750138,
      "num_tokens": 189013028.0,
      "step": 2630
    },
    {
      "entropy": 0.9832702308893204,
      "epoch": 0.7195421095666394,
      "grad_norm": 0.16413764655590057,
      "learning_rate": 9.589389534883721e-05,
      "loss": 1.4179,
      "mean_token_accuracy": 0.7364308670163154,
      "num_tokens": 189735349.0,
      "step": 2640
    },
    {
      "entropy": 0.9838227599859237,
      "epoch": 0.7222676478604524,
      "grad_norm": 0.18721725046634674,
      "learning_rate": 9.625726744186047e-05,
      "loss": 1.4213,
      "mean_token_accuracy": 0.7334291443228722,
      "num_tokens": 190432817.0,
      "step": 2650
    },
    {
      "entropy": 0.9814351618289947,
      "epoch": 0.7249931861542654,
      "grad_norm": 0.2767001688480377,
      "learning_rate": 9.662063953488373e-05,
      "loss": 1.4214,
      "mean_token_accuracy": 0.7366883307695389,
      "num_tokens": 191152964.0,
      "step": 2660
    },
    {
      "entropy": 0.9793224886059761,
      "epoch": 0.7277187244480785,
      "grad_norm": 0.1260397583246231,
      "learning_rate": 9.698401162790698e-05,
      "loss": 1.4216,
      "mean_token_accuracy": 0.7355253770947456,
      "num_tokens": 191864985.0,
      "step": 2670
    },
    {
      "entropy": 0.9775992184877396,
      "epoch": 0.7304442627418916,
      "grad_norm": 0.2061127871274948,
      "learning_rate": 9.734738372093024e-05,
      "loss": 1.42,
      "mean_token_accuracy": 0.7358565747737884,
      "num_tokens": 192582457.0,
      "step": 2680
    },
    {
      "entropy": 0.9830930635333062,
      "epoch": 0.7331698010357045,
      "grad_norm": 0.3281489610671997,
      "learning_rate": 9.77107558139535e-05,
      "loss": 1.4132,
      "mean_token_accuracy": 0.7400453820824623,
      "num_tokens": 193317939.0,
      "step": 2690
    },
    {
      "entropy": 0.9811283275485039,
      "epoch": 0.7358953393295176,
      "grad_norm": 0.18330809473991394,
      "learning_rate": 9.807412790697675e-05,
      "loss": 1.4198,
      "mean_token_accuracy": 0.7413364306092263,
      "num_tokens": 194045484.0,
      "step": 2700
    },
    {
      "entropy": 0.9767220064997673,
      "epoch": 0.7386208776233306,
      "grad_norm": 0.14695248007774353,
      "learning_rate": 9.84375e-05,
      "loss": 1.4279,
      "mean_token_accuracy": 0.7368486255407334,
      "num_tokens": 194750844.0,
      "step": 2710
    },
    {
      "entropy": 0.9996407404541969,
      "epoch": 0.7413464159171437,
      "grad_norm": 0.612859845161438,
      "learning_rate": 9.880087209302326e-05,
      "loss": 1.4393,
      "mean_token_accuracy": 0.7321041479706765,
      "num_tokens": 195469343.0,
      "step": 2720
    },
    {
      "entropy": 0.9824164986610413,
      "epoch": 0.7440719542109566,
      "grad_norm": 0.14814876019954681,
      "learning_rate": 9.916424418604652e-05,
      "loss": 1.4261,
      "mean_token_accuracy": 0.7364200308918953,
      "num_tokens": 196180730.0,
      "step": 2730
    },
    {
      "entropy": 0.9753024950623512,
      "epoch": 0.7467974925047697,
      "grad_norm": 0.3339892029762268,
      "learning_rate": 9.952761627906977e-05,
      "loss": 1.4118,
      "mean_token_accuracy": 0.7415054902434349,
      "num_tokens": 196917912.0,
      "step": 2740
    },
    {
      "entropy": 0.9833664178848267,
      "epoch": 0.7495230307985827,
      "grad_norm": 0.20620104670524597,
      "learning_rate": 9.989098837209303e-05,
      "loss": 1.4262,
      "mean_token_accuracy": 0.7400285750627518,
      "num_tokens": 197653923.0,
      "step": 2750
    },
    {
      "entropy": 0.989758312702179,
      "epoch": 0.7522485690923958,
      "grad_norm": 0.22599183022975922,
      "learning_rate": 9.99999955770215e-05,
      "loss": 1.4285,
      "mean_token_accuracy": 0.7386189118027687,
      "num_tokens": 198371819.0,
      "step": 2760
    },
    {
      "entropy": 0.9875634372234344,
      "epoch": 0.7549741073862087,
      "grad_norm": 0.1541522741317749,
      "learning_rate": 9.999997391345519e-05,
      "loss": 1.4323,
      "mean_token_accuracy": 0.7380873292684555,
      "num_tokens": 199082702.0,
      "step": 2770
    },
    {
      "entropy": 0.9855502024292946,
      "epoch": 0.7576996456800218,
      "grad_norm": 0.22785435616970062,
      "learning_rate": 9.999993419692509e-05,
      "loss": 1.4248,
      "mean_token_accuracy": 0.7359795659780503,
      "num_tokens": 199803798.0,
      "step": 2780
    },
    {
      "entropy": 0.9827453017234802,
      "epoch": 0.7604251839738349,
      "grad_norm": 0.15170662105083466,
      "learning_rate": 9.999987642744555e-05,
      "loss": 1.4129,
      "mean_token_accuracy": 0.7355940371751786,
      "num_tokens": 200523338.0,
      "step": 2790
    },
    {
      "entropy": 0.9769548758864403,
      "epoch": 0.7631507222676479,
      "grad_norm": 0.12554454803466797,
      "learning_rate": 9.99998006050374e-05,
      "loss": 1.427,
      "mean_token_accuracy": 0.7395160034298897,
      "num_tokens": 201243243.0,
      "step": 2800
    },
    {
      "entropy": 0.9889974981546402,
      "epoch": 0.7658762605614609,
      "grad_norm": 0.26222726702690125,
      "learning_rate": 9.999970672972803e-05,
      "loss": 1.422,
      "mean_token_accuracy": 0.736335052549839,
      "num_tokens": 201958044.0,
      "step": 2810
    },
    {
      "entropy": 0.9788802176713943,
      "epoch": 0.7686017988552739,
      "grad_norm": 0.19049885869026184,
      "learning_rate": 9.999959480155133e-05,
      "loss": 1.4169,
      "mean_token_accuracy": 0.7363810062408447,
      "num_tokens": 202664706.0,
      "step": 2820
    },
    {
      "entropy": 0.9821498796343804,
      "epoch": 0.771327337149087,
      "grad_norm": 0.161302387714386,
      "learning_rate": 9.999946482054772e-05,
      "loss": 1.4135,
      "mean_token_accuracy": 0.7354164943099022,
      "num_tokens": 203377268.0,
      "step": 2830
    },
    {
      "entropy": 0.9805738747119903,
      "epoch": 0.7740528754429,
      "grad_norm": 0.14993034303188324,
      "learning_rate": 9.999931678676413e-05,
      "loss": 1.4167,
      "mean_token_accuracy": 0.7371671542525291,
      "num_tokens": 204103306.0,
      "step": 2840
    },
    {
      "entropy": 0.98235704600811,
      "epoch": 0.776778413736713,
      "grad_norm": 0.22030942142009735,
      "learning_rate": 9.999915070025401e-05,
      "loss": 1.4179,
      "mean_token_accuracy": 0.7383449390530586,
      "num_tokens": 204825853.0,
      "step": 2850
    },
    {
      "entropy": 0.9786477074027061,
      "epoch": 0.779503952030526,
      "grad_norm": 0.15327267348766327,
      "learning_rate": 9.999896656107732e-05,
      "loss": 1.4213,
      "mean_token_accuracy": 0.7358534723520279,
      "num_tokens": 205535796.0,
      "step": 2860
    },
    {
      "entropy": 0.9758807629346847,
      "epoch": 0.7822294903243391,
      "grad_norm": 0.1187690794467926,
      "learning_rate": 9.999876436930056e-05,
      "loss": 1.4095,
      "mean_token_accuracy": 0.7400936737656594,
      "num_tokens": 206259558.0,
      "step": 2870
    },
    {
      "entropy": 0.9861981242895126,
      "epoch": 0.7849550286181521,
      "grad_norm": 0.13627082109451294,
      "learning_rate": 9.999854412499671e-05,
      "loss": 1.4272,
      "mean_token_accuracy": 0.7323023781180382,
      "num_tokens": 206970708.0,
      "step": 2880
    },
    {
      "entropy": 0.9836017206311226,
      "epoch": 0.7876805669119651,
      "grad_norm": 0.1861530840396881,
      "learning_rate": 9.999830582824532e-05,
      "loss": 1.4224,
      "mean_token_accuracy": 0.7317949622869492,
      "num_tokens": 207679412.0,
      "step": 2890
    },
    {
      "entropy": 0.9812063500285149,
      "epoch": 0.7904061052057781,
      "grad_norm": 0.2150312066078186,
      "learning_rate": 9.999804947913241e-05,
      "loss": 1.4233,
      "mean_token_accuracy": 0.740932235121727,
      "num_tokens": 208405196.0,
      "step": 2900
    },
    {
      "entropy": 0.9819123685359955,
      "epoch": 0.7931316434995912,
      "grad_norm": 0.5197040438652039,
      "learning_rate": 9.999777507775053e-05,
      "loss": 1.4179,
      "mean_token_accuracy": 0.7384748071432113,
      "num_tokens": 209134642.0,
      "step": 2910
    },
    {
      "entropy": 0.9714388608932495,
      "epoch": 0.7958571817934041,
      "grad_norm": 0.14457683265209198,
      "learning_rate": 9.999748262419877e-05,
      "loss": 1.4181,
      "mean_token_accuracy": 0.7427984416484833,
      "num_tokens": 209852921.0,
      "step": 2920
    },
    {
      "entropy": 0.9837919026613235,
      "epoch": 0.7985827200872172,
      "grad_norm": 0.18511174619197845,
      "learning_rate": 9.999717211858272e-05,
      "loss": 1.4216,
      "mean_token_accuracy": 0.7332604169845581,
      "num_tokens": 210559025.0,
      "step": 2930
    },
    {
      "entropy": 0.9797424361109733,
      "epoch": 0.8013082583810303,
      "grad_norm": 0.13574282824993134,
      "learning_rate": 9.999684356101448e-05,
      "loss": 1.4209,
      "mean_token_accuracy": 0.7408635228872299,
      "num_tokens": 211276251.0,
      "step": 2940
    },
    {
      "entropy": 0.9871427044272423,
      "epoch": 0.8040337966748433,
      "grad_norm": 0.1565672755241394,
      "learning_rate": 9.999649695161271e-05,
      "loss": 1.4225,
      "mean_token_accuracy": 0.734562648832798,
      "num_tokens": 211993021.0,
      "step": 2950
    },
    {
      "entropy": 0.984888119995594,
      "epoch": 0.8067593349686563,
      "grad_norm": 0.14846067130565643,
      "learning_rate": 9.999613229050252e-05,
      "loss": 1.4228,
      "mean_token_accuracy": 0.7389271706342697,
      "num_tokens": 212718426.0,
      "step": 2960
    },
    {
      "entropy": 0.9810102090239525,
      "epoch": 0.8094848732624693,
      "grad_norm": 0.14860044419765472,
      "learning_rate": 9.999574957781558e-05,
      "loss": 1.4162,
      "mean_token_accuracy": 0.7347591191530227,
      "num_tokens": 213438304.0,
      "step": 2970
    },
    {
      "entropy": 0.9810490146279335,
      "epoch": 0.8122104115562824,
      "grad_norm": 0.19743232429027557,
      "learning_rate": 9.999534881369011e-05,
      "loss": 1.4206,
      "mean_token_accuracy": 0.739751273393631,
      "num_tokens": 214165613.0,
      "step": 2980
    },
    {
      "entropy": 0.9807188436388969,
      "epoch": 0.8149359498500954,
      "grad_norm": 0.1587584763765335,
      "learning_rate": 9.999492999827077e-05,
      "loss": 1.4276,
      "mean_token_accuracy": 0.7378657341003418,
      "num_tokens": 214876982.0,
      "step": 2990
    },
    {
      "entropy": 0.9837844550609589,
      "epoch": 0.8176614881439084,
      "grad_norm": 0.11708062142133713,
      "learning_rate": 9.999449313170879e-05,
      "loss": 1.4225,
      "mean_token_accuracy": 0.7384473770856858,
      "num_tokens": 215594732.0,
      "step": 3000
    },
    {
      "entropy": 0.9751689925789833,
      "epoch": 0.8203870264377214,
      "grad_norm": 0.14906936883926392,
      "learning_rate": 9.99940382141619e-05,
      "loss": 1.4111,
      "mean_token_accuracy": 0.736640265583992,
      "num_tokens": 216309994.0,
      "step": 3010
    },
    {
      "entropy": 0.974778625369072,
      "epoch": 0.8231125647315345,
      "grad_norm": 0.20057159662246704,
      "learning_rate": 9.999356524579435e-05,
      "loss": 1.4122,
      "mean_token_accuracy": 0.7381565615534782,
      "num_tokens": 217017050.0,
      "step": 3020
    },
    {
      "entropy": 0.9723775506019592,
      "epoch": 0.8258381030253475,
      "grad_norm": 0.2575012445449829,
      "learning_rate": 9.999307422677692e-05,
      "loss": 1.4062,
      "mean_token_accuracy": 0.7382101267576218,
      "num_tokens": 217738934.0,
      "step": 3030
    },
    {
      "entropy": 0.9816832140088081,
      "epoch": 0.8285636413191605,
      "grad_norm": 0.13900838792324066,
      "learning_rate": 9.99925651572869e-05,
      "loss": 1.4224,
      "mean_token_accuracy": 0.7348329618573188,
      "num_tokens": 218453524.0,
      "step": 3040
    },
    {
      "entropy": 0.982389472424984,
      "epoch": 0.8312891796129735,
      "grad_norm": 0.15657106041908264,
      "learning_rate": 9.99920380375081e-05,
      "loss": 1.4253,
      "mean_token_accuracy": 0.7420304551720619,
      "num_tokens": 219187154.0,
      "step": 3050
    },
    {
      "entropy": 0.9812181010842324,
      "epoch": 0.8340147179067866,
      "grad_norm": 0.189975768327713,
      "learning_rate": 9.99914928676308e-05,
      "loss": 1.4113,
      "mean_token_accuracy": 0.7373171448707581,
      "num_tokens": 219910460.0,
      "step": 3060
    },
    {
      "entropy": 0.9803787097334862,
      "epoch": 0.8367402562005997,
      "grad_norm": 0.14406999945640564,
      "learning_rate": 9.999092964785189e-05,
      "loss": 1.4144,
      "mean_token_accuracy": 0.7389702558517456,
      "num_tokens": 220638772.0,
      "step": 3070
    },
    {
      "entropy": 0.978831271827221,
      "epoch": 0.8394657944944126,
      "grad_norm": 0.12598539888858795,
      "learning_rate": 9.999034837837469e-05,
      "loss": 1.4097,
      "mean_token_accuracy": 0.7423653185367585,
      "num_tokens": 221364839.0,
      "step": 3080
    },
    {
      "entropy": 0.977126270532608,
      "epoch": 0.8421913327882257,
      "grad_norm": 0.1539270281791687,
      "learning_rate": 9.998974905940907e-05,
      "loss": 1.4135,
      "mean_token_accuracy": 0.7419637933373451,
      "num_tokens": 222096513.0,
      "step": 3090
    },
    {
      "entropy": 0.9715770184993744,
      "epoch": 0.8449168710820387,
      "grad_norm": 0.16594548523426056,
      "learning_rate": 9.998913169117146e-05,
      "loss": 1.4097,
      "mean_token_accuracy": 0.7437430068850517,
      "num_tokens": 222826039.0,
      "step": 3100
    },
    {
      "entropy": 0.9744135066866875,
      "epoch": 0.8476424093758518,
      "grad_norm": 0.12864826619625092,
      "learning_rate": 9.998849627388473e-05,
      "loss": 1.409,
      "mean_token_accuracy": 0.7474918752908707,
      "num_tokens": 223570545.0,
      "step": 3110
    },
    {
      "entropy": 0.9805658519268036,
      "epoch": 0.8503679476696647,
      "grad_norm": 0.1582091897726059,
      "learning_rate": 9.998784280777834e-05,
      "loss": 1.4142,
      "mean_token_accuracy": 0.7373319238424301,
      "num_tokens": 224288940.0,
      "step": 3120
    },
    {
      "entropy": 0.974412453174591,
      "epoch": 0.8530934859634778,
      "grad_norm": 0.1336938887834549,
      "learning_rate": 9.998717129308818e-05,
      "loss": 1.4103,
      "mean_token_accuracy": 0.7408842518925667,
      "num_tokens": 225023171.0,
      "step": 3130
    },
    {
      "entropy": 0.9713442757725715,
      "epoch": 0.8558190242572908,
      "grad_norm": 0.10065404325723648,
      "learning_rate": 9.998648173005674e-05,
      "loss": 1.4072,
      "mean_token_accuracy": 0.7376281142234802,
      "num_tokens": 225729545.0,
      "step": 3140
    },
    {
      "entropy": 0.9841649621725083,
      "epoch": 0.8585445625511039,
      "grad_norm": 0.15625248849391937,
      "learning_rate": 9.998577411893299e-05,
      "loss": 1.4234,
      "mean_token_accuracy": 0.7346471220254898,
      "num_tokens": 226444810.0,
      "step": 3150
    },
    {
      "entropy": 0.9743556290864944,
      "epoch": 0.8612701008449168,
      "grad_norm": 0.15847258269786835,
      "learning_rate": 9.99850484599724e-05,
      "loss": 1.4183,
      "mean_token_accuracy": 0.7372889712452888,
      "num_tokens": 227162826.0,
      "step": 3160
    },
    {
      "entropy": 0.9804324865341186,
      "epoch": 0.8639956391387299,
      "grad_norm": 0.12864305078983307,
      "learning_rate": 9.998430475343701e-05,
      "loss": 1.4199,
      "mean_token_accuracy": 0.7350202083587647,
      "num_tokens": 227869180.0,
      "step": 3170
    },
    {
      "entropy": 0.9829505577683448,
      "epoch": 0.866721177432543,
      "grad_norm": 0.14491578936576843,
      "learning_rate": 9.998354299959534e-05,
      "loss": 1.4187,
      "mean_token_accuracy": 0.7396857485175132,
      "num_tokens": 228596762.0,
      "step": 3180
    },
    {
      "entropy": 0.9742689624428749,
      "epoch": 0.8694467157263559,
      "grad_norm": 0.1269872784614563,
      "learning_rate": 9.998276319872238e-05,
      "loss": 1.4101,
      "mean_token_accuracy": 0.7371434390544891,
      "num_tokens": 229303562.0,
      "step": 3190
    },
    {
      "entropy": 0.9751578256487846,
      "epoch": 0.872172254020169,
      "grad_norm": 0.21285098791122437,
      "learning_rate": 9.998196535109973e-05,
      "loss": 1.4199,
      "mean_token_accuracy": 0.736963638663292,
      "num_tokens": 230007345.0,
      "step": 3200
    },
    {
      "entropy": 0.9789491593837738,
      "epoch": 0.874897792313982,
      "grad_norm": 0.26488471031188965,
      "learning_rate": 9.998114945701544e-05,
      "loss": 1.4277,
      "mean_token_accuracy": 0.735780592262745,
      "num_tokens": 230705545.0,
      "step": 3210
    },
    {
      "entropy": 0.9745857268571854,
      "epoch": 0.8776233306077951,
      "grad_norm": 0.1784295290708542,
      "learning_rate": 9.998031551676413e-05,
      "loss": 1.412,
      "mean_token_accuracy": 0.7361106410622597,
      "num_tokens": 231420484.0,
      "step": 3220
    },
    {
      "entropy": 0.993267172574997,
      "epoch": 0.880348868901608,
      "grad_norm": 0.17070242762565613,
      "learning_rate": 9.997946353064685e-05,
      "loss": 1.4232,
      "mean_token_accuracy": 0.7339597791433334,
      "num_tokens": 232141626.0,
      "step": 3230
    },
    {
      "entropy": 0.9836523786187172,
      "epoch": 0.8830744071954211,
      "grad_norm": 0.15658451616764069,
      "learning_rate": 9.997859349897125e-05,
      "loss": 1.4325,
      "mean_token_accuracy": 0.737040302157402,
      "num_tokens": 232854333.0,
      "step": 3240
    },
    {
      "entropy": 0.9760024771094322,
      "epoch": 0.8857999454892341,
      "grad_norm": 0.19112402200698853,
      "learning_rate": 9.997770542205144e-05,
      "loss": 1.4119,
      "mean_token_accuracy": 0.7414720132946968,
      "num_tokens": 233590794.0,
      "step": 3250
    },
    {
      "entropy": 0.9777669444680214,
      "epoch": 0.8885254837830472,
      "grad_norm": 0.11839189380407333,
      "learning_rate": 9.997679930020811e-05,
      "loss": 1.4153,
      "mean_token_accuracy": 0.734228539466858,
      "num_tokens": 234295557.0,
      "step": 3260
    },
    {
      "entropy": 0.9728788286447525,
      "epoch": 0.8912510220768601,
      "grad_norm": 0.10589392483234406,
      "learning_rate": 9.99758751337684e-05,
      "loss": 1.4056,
      "mean_token_accuracy": 0.7415377557277679,
      "num_tokens": 235025124.0,
      "step": 3270
    },
    {
      "entropy": 0.971225069463253,
      "epoch": 0.8939765603706732,
      "grad_norm": 0.21799670159816742,
      "learning_rate": 9.997493292306596e-05,
      "loss": 1.4112,
      "mean_token_accuracy": 0.7429666057229042,
      "num_tokens": 235760710.0,
      "step": 3280
    },
    {
      "entropy": 0.9753315597772598,
      "epoch": 0.8967020986644862,
      "grad_norm": 0.19039765000343323,
      "learning_rate": 9.997397266844104e-05,
      "loss": 1.4069,
      "mean_token_accuracy": 0.7381700858473778,
      "num_tokens": 236483738.0,
      "step": 3290
    },
    {
      "entropy": 0.9832386106252671,
      "epoch": 0.8994276369582993,
      "grad_norm": 0.13938450813293457,
      "learning_rate": 9.997299437024032e-05,
      "loss": 1.4236,
      "mean_token_accuracy": 0.7357633501291275,
      "num_tokens": 237194525.0,
      "step": 3300
    },
    {
      "entropy": 0.9810293108224869,
      "epoch": 0.9021531752521122,
      "grad_norm": 0.17831286787986755,
      "learning_rate": 9.9971998028817e-05,
      "loss": 1.4156,
      "mean_token_accuracy": 0.7319070100784302,
      "num_tokens": 237900874.0,
      "step": 3310
    },
    {
      "entropy": 0.9783111020922661,
      "epoch": 0.9048787135459253,
      "grad_norm": 0.1969386339187622,
      "learning_rate": 9.997098364453087e-05,
      "loss": 1.4194,
      "mean_token_accuracy": 0.7367761194705963,
      "num_tokens": 238608497.0,
      "step": 3320
    },
    {
      "entropy": 0.9764347687363625,
      "epoch": 0.9076042518397384,
      "grad_norm": 0.14776062965393066,
      "learning_rate": 9.996995121774814e-05,
      "loss": 1.4232,
      "mean_token_accuracy": 0.7391286328434944,
      "num_tokens": 239326913.0,
      "step": 3330
    },
    {
      "entropy": 0.9870787173509598,
      "epoch": 0.9103297901335514,
      "grad_norm": 0.14292185008525848,
      "learning_rate": 9.99689007488416e-05,
      "loss": 1.4188,
      "mean_token_accuracy": 0.7383605018258095,
      "num_tokens": 240060814.0,
      "step": 3340
    },
    {
      "entropy": 0.978061044216156,
      "epoch": 0.9130553284273644,
      "grad_norm": 0.14178350567817688,
      "learning_rate": 9.996783223819053e-05,
      "loss": 1.4168,
      "mean_token_accuracy": 0.7363059341907501,
      "num_tokens": 240779526.0,
      "step": 3350
    },
    {
      "entropy": 0.9764518067240715,
      "epoch": 0.9157808667211774,
      "grad_norm": 0.272127628326416,
      "learning_rate": 9.996674568618072e-05,
      "loss": 1.4192,
      "mean_token_accuracy": 0.7388992846012116,
      "num_tokens": 241495307.0,
      "step": 3360
    },
    {
      "entropy": 0.974106739461422,
      "epoch": 0.9185064050149905,
      "grad_norm": 0.1740814596414566,
      "learning_rate": 9.996564109320448e-05,
      "loss": 1.4103,
      "mean_token_accuracy": 0.7449988663196564,
      "num_tokens": 242230783.0,
      "step": 3370
    },
    {
      "entropy": 0.9787409767508507,
      "epoch": 0.9212319433088035,
      "grad_norm": 0.10566773265600204,
      "learning_rate": 9.996451845966064e-05,
      "loss": 1.4163,
      "mean_token_accuracy": 0.7380370870232582,
      "num_tokens": 242940041.0,
      "step": 3380
    },
    {
      "entropy": 0.9847319096326828,
      "epoch": 0.9239574816026165,
      "grad_norm": 0.10505988448858261,
      "learning_rate": 9.996337778595453e-05,
      "loss": 1.4133,
      "mean_token_accuracy": 0.7361453324556351,
      "num_tokens": 243665358.0,
      "step": 3390
    },
    {
      "entropy": 0.9767900019884109,
      "epoch": 0.9266830198964295,
      "grad_norm": 0.12417960166931152,
      "learning_rate": 9.996221907249799e-05,
      "loss": 1.4147,
      "mean_token_accuracy": 0.7380743145942688,
      "num_tokens": 244394333.0,
      "step": 3400
    },
    {
      "entropy": 0.9792727619409561,
      "epoch": 0.9294085581902426,
      "grad_norm": 0.11050882935523987,
      "learning_rate": 9.996104231970942e-05,
      "loss": 1.4155,
      "mean_token_accuracy": 0.7417500883340835,
      "num_tokens": 245124365.0,
      "step": 3410
    },
    {
      "entropy": 0.9754124477505683,
      "epoch": 0.9321340964840557,
      "grad_norm": 0.11477773636579514,
      "learning_rate": 9.995984752801367e-05,
      "loss": 1.4135,
      "mean_token_accuracy": 0.7388074770569801,
      "num_tokens": 245835476.0,
      "step": 3420
    },
    {
      "entropy": 0.9770966023206711,
      "epoch": 0.9348596347778686,
      "grad_norm": 0.33463749289512634,
      "learning_rate": 9.995863469784213e-05,
      "loss": 1.4192,
      "mean_token_accuracy": 0.7341749802231788,
      "num_tokens": 246545852.0,
      "step": 3430
    },
    {
      "entropy": 0.9832690536975861,
      "epoch": 0.9375851730716817,
      "grad_norm": 0.13229504227638245,
      "learning_rate": 9.995740382963272e-05,
      "loss": 1.4156,
      "mean_token_accuracy": 0.7382375612854958,
      "num_tokens": 247261964.0,
      "step": 3440
    },
    {
      "entropy": 0.9732533872127533,
      "epoch": 0.9403107113654947,
      "grad_norm": 0.3632689118385315,
      "learning_rate": 9.995615492382983e-05,
      "loss": 1.4205,
      "mean_token_accuracy": 0.7353919863700866,
      "num_tokens": 247959080.0,
      "step": 3450
    },
    {
      "entropy": 0.9803975969552994,
      "epoch": 0.9430362496593078,
      "grad_norm": 0.16551385819911957,
      "learning_rate": 9.995488798088442e-05,
      "loss": 1.4225,
      "mean_token_accuracy": 0.7330883577466011,
      "num_tokens": 248657170.0,
      "step": 3460
    },
    {
      "entropy": 0.971314737200737,
      "epoch": 0.9457617879531207,
      "grad_norm": 0.10738106071949005,
      "learning_rate": 9.995360300125392e-05,
      "loss": 1.4166,
      "mean_token_accuracy": 0.7408865720033646,
      "num_tokens": 249385976.0,
      "step": 3470
    },
    {
      "entropy": 0.9782015651464462,
      "epoch": 0.9484873262469338,
      "grad_norm": 0.1215008944272995,
      "learning_rate": 9.995229998540228e-05,
      "loss": 1.4229,
      "mean_token_accuracy": 0.7396703884005547,
      "num_tokens": 250105086.0,
      "step": 3480
    },
    {
      "entropy": 0.9830413952469825,
      "epoch": 0.9512128645407468,
      "grad_norm": 0.11202839761972427,
      "learning_rate": 9.995097893379997e-05,
      "loss": 1.4209,
      "mean_token_accuracy": 0.7345697537064553,
      "num_tokens": 250830736.0,
      "step": 3490
    },
    {
      "entropy": 0.9833935052156448,
      "epoch": 0.9539384028345598,
      "grad_norm": 0.12974119186401367,
      "learning_rate": 9.994963984692395e-05,
      "loss": 1.4154,
      "mean_token_accuracy": 0.7403559491038323,
      "num_tokens": 251562656.0,
      "step": 3500
    },
    {
      "entropy": 0.9744713872671127,
      "epoch": 0.9566639411283728,
      "grad_norm": 0.13219153881072998,
      "learning_rate": 9.994828272525774e-05,
      "loss": 1.4123,
      "mean_token_accuracy": 0.7411713972687721,
      "num_tokens": 252286064.0,
      "step": 3510
    },
    {
      "entropy": 0.9831308081746102,
      "epoch": 0.9593894794221859,
      "grad_norm": 0.14877712726593018,
      "learning_rate": 9.994690756929132e-05,
      "loss": 1.4207,
      "mean_token_accuracy": 0.7395951136946678,
      "num_tokens": 252998169.0,
      "step": 3520
    },
    {
      "entropy": 0.9759589090943337,
      "epoch": 0.9621150177159989,
      "grad_norm": 0.17554226517677307,
      "learning_rate": 9.994551437952123e-05,
      "loss": 1.4155,
      "mean_token_accuracy": 0.7382798790931702,
      "num_tokens": 253721062.0,
      "step": 3530
    },
    {
      "entropy": 0.9726287558674812,
      "epoch": 0.9648405560098119,
      "grad_norm": 0.12672601640224457,
      "learning_rate": 9.994410315645047e-05,
      "loss": 1.4139,
      "mean_token_accuracy": 0.7368972420692443,
      "num_tokens": 254433375.0,
      "step": 3540
    },
    {
      "entropy": 0.9700451165437698,
      "epoch": 0.967566094303625,
      "grad_norm": 0.1541493982076645,
      "learning_rate": 9.994267390058858e-05,
      "loss": 1.4172,
      "mean_token_accuracy": 0.7434371784329414,
      "num_tokens": 255160255.0,
      "step": 3550
    },
    {
      "entropy": 0.9721565470099449,
      "epoch": 0.970291632597438,
      "grad_norm": 0.15291433036327362,
      "learning_rate": 9.994122661245159e-05,
      "loss": 1.4109,
      "mean_token_accuracy": 0.7407122254371643,
      "num_tokens": 255883769.0,
      "step": 3560
    },
    {
      "entropy": 0.9811560899019242,
      "epoch": 0.9730171708912511,
      "grad_norm": 0.10619845241308212,
      "learning_rate": 9.993976129256209e-05,
      "loss": 1.4204,
      "mean_token_accuracy": 0.7353764981031418,
      "num_tokens": 256598991.0,
      "step": 3570
    },
    {
      "entropy": 0.9756833255290985,
      "epoch": 0.975742709185064,
      "grad_norm": 0.12091992795467377,
      "learning_rate": 9.993827794144914e-05,
      "loss": 1.419,
      "mean_token_accuracy": 0.7370814487338067,
      "num_tokens": 257306705.0,
      "step": 3580
    },
    {
      "entropy": 0.9844296738505364,
      "epoch": 0.9784682474788771,
      "grad_norm": 0.12438544631004333,
      "learning_rate": 9.99367765596483e-05,
      "loss": 1.4184,
      "mean_token_accuracy": 0.7343698784708976,
      "num_tokens": 258013773.0,
      "step": 3590
    },
    {
      "entropy": 0.9808205753564835,
      "epoch": 0.9811937857726901,
      "grad_norm": 0.1304776519536972,
      "learning_rate": 9.993525714770166e-05,
      "loss": 1.413,
      "mean_token_accuracy": 0.736425158381462,
      "num_tokens": 258731029.0,
      "step": 3600
    },
    {
      "entropy": 0.9760351404547691,
      "epoch": 0.9839193240665032,
      "grad_norm": 0.1096147745847702,
      "learning_rate": 9.993371970615785e-05,
      "loss": 1.4177,
      "mean_token_accuracy": 0.7411983251571655,
      "num_tokens": 259459979.0,
      "step": 3610
    },
    {
      "entropy": 0.9712781876325607,
      "epoch": 0.9866448623603161,
      "grad_norm": 0.14058972895145416,
      "learning_rate": 9.993216423557194e-05,
      "loss": 1.4085,
      "mean_token_accuracy": 0.7394808471202851,
      "num_tokens": 260180085.0,
      "step": 3620
    },
    {
      "entropy": 0.9717079535126686,
      "epoch": 0.9893704006541292,
      "grad_norm": 0.11068553477525711,
      "learning_rate": 9.993059073650556e-05,
      "loss": 1.4056,
      "mean_token_accuracy": 0.7396122291684151,
      "num_tokens": 260907071.0,
      "step": 3630
    },
    {
      "entropy": 0.9745209664106369,
      "epoch": 0.9920959389479422,
      "grad_norm": 0.12133318185806274,
      "learning_rate": 9.992899920952683e-05,
      "loss": 1.4106,
      "mean_token_accuracy": 0.7415842488408089,
      "num_tokens": 261639149.0,
      "step": 3640
    },
    {
      "entropy": 0.9628227591514588,
      "epoch": 0.9948214772417553,
      "grad_norm": 0.3034980297088623,
      "learning_rate": 9.992738965521041e-05,
      "loss": 1.4096,
      "mean_token_accuracy": 0.7393161743879318,
      "num_tokens": 262346251.0,
      "step": 3650
    },
    {
      "entropy": 0.9788528263568879,
      "epoch": 0.9975470155355682,
      "grad_norm": 0.12548959255218506,
      "learning_rate": 9.992576207413741e-05,
      "loss": 1.4125,
      "mean_token_accuracy": 0.7384046420454979,
      "num_tokens": 263065815.0,
      "step": 3660
    },
    {
      "entropy": 0.9772783532738686,
      "epoch": 1.0002725538293813,
      "grad_norm": 0.13550417125225067,
      "learning_rate": 9.992411646689552e-05,
      "loss": 1.4132,
      "mean_token_accuracy": 0.7365165814757347,
      "num_tokens": 263772518.0,
      "step": 3670
    },
    {
      "entropy": 0.97506842315197,
      "epoch": 1.0029980921231942,
      "grad_norm": 0.12043224275112152,
      "learning_rate": 9.992245283407888e-05,
      "loss": 1.4098,
      "mean_token_accuracy": 0.7382272049784661,
      "num_tokens": 264486665.0,
      "step": 3680
    },
    {
      "entropy": 0.9778215289115906,
      "epoch": 1.0057236304170074,
      "grad_norm": 0.12420505285263062,
      "learning_rate": 9.992077117628817e-05,
      "loss": 1.4161,
      "mean_token_accuracy": 0.7353036567568779,
      "num_tokens": 265187611.0,
      "step": 3690
    },
    {
      "entropy": 0.969636894762516,
      "epoch": 1.0084491687108204,
      "grad_norm": 0.1205442026257515,
      "learning_rate": 9.991907149413055e-05,
      "loss": 1.411,
      "mean_token_accuracy": 0.7377049684524536,
      "num_tokens": 265896405.0,
      "step": 3700
    },
    {
      "entropy": 0.9737311184406281,
      "epoch": 1.0111747070046335,
      "grad_norm": 0.11289207637310028,
      "learning_rate": 9.991735378821973e-05,
      "loss": 1.4059,
      "mean_token_accuracy": 0.7394744753837585,
      "num_tokens": 266622529.0,
      "step": 3710
    },
    {
      "entropy": 0.9757361307740211,
      "epoch": 1.0139002452984465,
      "grad_norm": 0.12090495973825455,
      "learning_rate": 9.991561805917591e-05,
      "loss": 1.4122,
      "mean_token_accuracy": 0.7386458143591881,
      "num_tokens": 267335050.0,
      "step": 3720
    },
    {
      "entropy": 0.964876189827919,
      "epoch": 1.0166257835922594,
      "grad_norm": 0.26090943813323975,
      "learning_rate": 9.991386430762574e-05,
      "loss": 1.4113,
      "mean_token_accuracy": 0.7410697713494301,
      "num_tokens": 268058316.0,
      "step": 3730
    },
    {
      "entropy": 0.9755144536495208,
      "epoch": 1.0193513218860726,
      "grad_norm": 0.1461142748594284,
      "learning_rate": 9.99120925342025e-05,
      "loss": 1.4123,
      "mean_token_accuracy": 0.7343229576945305,
      "num_tokens": 268762648.0,
      "step": 3740
    },
    {
      "entropy": 0.9748295471072197,
      "epoch": 1.0220768601798855,
      "grad_norm": 0.11366649717092514,
      "learning_rate": 9.991030273954585e-05,
      "loss": 1.4121,
      "mean_token_accuracy": 0.7391687527298927,
      "num_tokens": 269489178.0,
      "step": 3750
    },
    {
      "entropy": 0.9809691607952118,
      "epoch": 1.0248023984736985,
      "grad_norm": 0.11171828955411911,
      "learning_rate": 9.990849492430203e-05,
      "loss": 1.411,
      "mean_token_accuracy": 0.7341507405042649,
      "num_tokens": 270199618.0,
      "step": 3760
    },
    {
      "entropy": 0.9676360934972763,
      "epoch": 1.0275279367675116,
      "grad_norm": 0.2335195094347,
      "learning_rate": 9.990666908912376e-05,
      "loss": 1.4024,
      "mean_token_accuracy": 0.7410795643925667,
      "num_tokens": 270911991.0,
      "step": 3770
    },
    {
      "entropy": 0.9800814092159271,
      "epoch": 1.0302534750613246,
      "grad_norm": 0.21199901401996613,
      "learning_rate": 9.990482523467031e-05,
      "loss": 1.417,
      "mean_token_accuracy": 0.7395343884825707,
      "num_tokens": 271632431.0,
      "step": 3780
    },
    {
      "entropy": 0.979529619216919,
      "epoch": 1.0329790133551375,
      "grad_norm": 0.15840569138526917,
      "learning_rate": 9.990296336160739e-05,
      "loss": 1.4118,
      "mean_token_accuracy": 0.7357161551713943,
      "num_tokens": 272349165.0,
      "step": 3790
    },
    {
      "entropy": 0.9744057133793831,
      "epoch": 1.0357045516489507,
      "grad_norm": 0.15760374069213867,
      "learning_rate": 9.990108347060723e-05,
      "loss": 1.416,
      "mean_token_accuracy": 0.7330457016825676,
      "num_tokens": 273045577.0,
      "step": 3800
    },
    {
      "entropy": 0.9722929179668427,
      "epoch": 1.0384300899427636,
      "grad_norm": 0.14145927131175995,
      "learning_rate": 9.989918556234864e-05,
      "loss": 1.4064,
      "mean_token_accuracy": 0.7408761292695999,
      "num_tokens": 273769311.0,
      "step": 3810
    },
    {
      "entropy": 0.9685794860124588,
      "epoch": 1.0411556282365768,
      "grad_norm": 0.1353922039270401,
      "learning_rate": 9.989726963751682e-05,
      "loss": 1.4061,
      "mean_token_accuracy": 0.7405291020870208,
      "num_tokens": 274490053.0,
      "step": 3820
    },
    {
      "entropy": 0.9852296590805054,
      "epoch": 1.0438811665303898,
      "grad_norm": 0.1262388378381729,
      "learning_rate": 9.989533569680356e-05,
      "loss": 1.4211,
      "mean_token_accuracy": 0.737436780333519,
      "num_tokens": 275211918.0,
      "step": 3830
    },
    {
      "entropy": 0.9729504108428955,
      "epoch": 1.0466067048242027,
      "grad_norm": 0.11301837861537933,
      "learning_rate": 9.989338374090713e-05,
      "loss": 1.4085,
      "mean_token_accuracy": 0.7369310811161995,
      "num_tokens": 275929181.0,
      "step": 3840
    },
    {
      "entropy": 0.9741118997335434,
      "epoch": 1.0493322431180159,
      "grad_norm": 0.13260625302791595,
      "learning_rate": 9.98914137705323e-05,
      "loss": 1.4143,
      "mean_token_accuracy": 0.7391919955611229,
      "num_tokens": 276653068.0,
      "step": 3850
    },
    {
      "entropy": 0.9720348879694939,
      "epoch": 1.0520577814118288,
      "grad_norm": 0.13643941283226013,
      "learning_rate": 9.988942578639034e-05,
      "loss": 1.4084,
      "mean_token_accuracy": 0.7384933397173882,
      "num_tokens": 277379115.0,
      "step": 3860
    },
    {
      "entropy": 0.9685726597905159,
      "epoch": 1.0547833197056418,
      "grad_norm": 0.14767953753471375,
      "learning_rate": 9.988741978919902e-05,
      "loss": 1.4046,
      "mean_token_accuracy": 0.7400298252701759,
      "num_tokens": 278095326.0,
      "step": 3870
    },
    {
      "entropy": 0.9742128312587738,
      "epoch": 1.057508857999455,
      "grad_norm": 0.10921667516231537,
      "learning_rate": 9.988539577968265e-05,
      "loss": 1.403,
      "mean_token_accuracy": 0.737097255885601,
      "num_tokens": 278815900.0,
      "step": 3880
    },
    {
      "entropy": 0.9724373623728753,
      "epoch": 1.0602343962932679,
      "grad_norm": 0.1891479641199112,
      "learning_rate": 9.9883353758572e-05,
      "loss": 1.4082,
      "mean_token_accuracy": 0.7374177858233452,
      "num_tokens": 279524148.0,
      "step": 3890
    },
    {
      "entropy": 0.9676162213087082,
      "epoch": 1.062959934587081,
      "grad_norm": 0.2643381655216217,
      "learning_rate": 9.988129372660437e-05,
      "loss": 1.4054,
      "mean_token_accuracy": 0.7386135622859001,
      "num_tokens": 280232785.0,
      "step": 3900
    },
    {
      "entropy": 0.9662006422877312,
      "epoch": 1.065685472880894,
      "grad_norm": 0.13425377011299133,
      "learning_rate": 9.987921568452356e-05,
      "loss": 1.4007,
      "mean_token_accuracy": 0.7396600037813187,
      "num_tokens": 280953294.0,
      "step": 3910
    },
    {
      "entropy": 0.9768855273723602,
      "epoch": 1.068411011174707,
      "grad_norm": 0.12102607637643814,
      "learning_rate": 9.987711963307984e-05,
      "loss": 1.408,
      "mean_token_accuracy": 0.7410584643483162,
      "num_tokens": 281684292.0,
      "step": 3920
    },
    {
      "entropy": 0.9696345388889313,
      "epoch": 1.07113654946852,
      "grad_norm": 0.12390506267547607,
      "learning_rate": 9.987500557303004e-05,
      "loss": 1.4078,
      "mean_token_accuracy": 0.7363270789384841,
      "num_tokens": 282397535.0,
      "step": 3930
    },
    {
      "entropy": 0.9763494119048118,
      "epoch": 1.073862087762333,
      "grad_norm": 0.15260037779808044,
      "learning_rate": 9.987287350513743e-05,
      "loss": 1.4212,
      "mean_token_accuracy": 0.7359601497650147,
      "num_tokens": 283106427.0,
      "step": 3940
    },
    {
      "entropy": 0.9774065673351288,
      "epoch": 1.076587626056146,
      "grad_norm": 0.16525544226169586,
      "learning_rate": 9.987072343017185e-05,
      "loss": 1.4055,
      "mean_token_accuracy": 0.7404107555747033,
      "num_tokens": 283837403.0,
      "step": 3950
    },
    {
      "entropy": 0.9703848645091057,
      "epoch": 1.0793131643499592,
      "grad_norm": 0.13805174827575684,
      "learning_rate": 9.986855534890957e-05,
      "loss": 1.4112,
      "mean_token_accuracy": 0.7389624550938606,
      "num_tokens": 284548634.0,
      "step": 3960
    },
    {
      "entropy": 0.9771598130464554,
      "epoch": 1.082038702643772,
      "grad_norm": 0.11810410022735596,
      "learning_rate": 9.986636926213341e-05,
      "loss": 1.4055,
      "mean_token_accuracy": 0.7352323740720749,
      "num_tokens": 285256203.0,
      "step": 3970
    },
    {
      "entropy": 0.9775662258267402,
      "epoch": 1.0847642409375853,
      "grad_norm": 0.11807934939861298,
      "learning_rate": 9.986416517063271e-05,
      "loss": 1.4211,
      "mean_token_accuracy": 0.7364842385053635,
      "num_tokens": 285961891.0,
      "step": 3980
    },
    {
      "entropy": 0.9697574928402901,
      "epoch": 1.0874897792313982,
      "grad_norm": 0.10679399967193604,
      "learning_rate": 9.986194307520323e-05,
      "loss": 1.4063,
      "mean_token_accuracy": 0.7410814076662063,
      "num_tokens": 286672644.0,
      "step": 3990
    },
    {
      "entropy": 0.967772725224495,
      "epoch": 1.0902153175252112,
      "grad_norm": 0.1722603440284729,
      "learning_rate": 9.985970297664728e-05,
      "loss": 1.4025,
      "mean_token_accuracy": 0.7408358335494996,
      "num_tokens": 287393348.0,
      "step": 4000
    },
    {
      "entropy": 0.967395193874836,
      "epoch": 1.0929408558190243,
      "grad_norm": 0.1754252314567566,
      "learning_rate": 9.98574448757737e-05,
      "loss": 1.4075,
      "mean_token_accuracy": 0.7433686286211014,
      "num_tokens": 288124634.0,
      "step": 4010
    },
    {
      "entropy": 0.9791459232568741,
      "epoch": 1.0956663941128373,
      "grad_norm": 0.13418108224868774,
      "learning_rate": 9.985516877339778e-05,
      "loss": 1.4096,
      "mean_token_accuracy": 0.7463141262531281,
      "num_tokens": 288875900.0,
      "step": 4020
    },
    {
      "entropy": 0.966758668422699,
      "epoch": 1.0983919324066502,
      "grad_norm": 0.10789959132671356,
      "learning_rate": 9.985287467034133e-05,
      "loss": 1.4133,
      "mean_token_accuracy": 0.7402540147304535,
      "num_tokens": 289579323.0,
      "step": 4030
    },
    {
      "entropy": 0.9771617695689201,
      "epoch": 1.1011174707004634,
      "grad_norm": 0.24271731078624725,
      "learning_rate": 9.985056256743266e-05,
      "loss": 1.4102,
      "mean_token_accuracy": 0.7404718190431595,
      "num_tokens": 290304508.0,
      "step": 4040
    },
    {
      "entropy": 0.9759747684001923,
      "epoch": 1.1038430089942763,
      "grad_norm": 0.13188596069812775,
      "learning_rate": 9.984823246550659e-05,
      "loss": 1.411,
      "mean_token_accuracy": 0.740820249915123,
      "num_tokens": 291030812.0,
      "step": 4050
    },
    {
      "entropy": 0.9823436632752418,
      "epoch": 1.1065685472880893,
      "grad_norm": 0.13854913413524628,
      "learning_rate": 9.984588436540438e-05,
      "loss": 1.4271,
      "mean_token_accuracy": 0.7301308572292328,
      "num_tokens": 291727243.0,
      "step": 4060
    },
    {
      "entropy": 0.9707790210843086,
      "epoch": 1.1092940855819025,
      "grad_norm": 0.15856264531612396,
      "learning_rate": 9.984351826797389e-05,
      "loss": 1.4069,
      "mean_token_accuracy": 0.7377354234457016,
      "num_tokens": 292445092.0,
      "step": 4070
    },
    {
      "entropy": 0.9699542596936226,
      "epoch": 1.1120196238757154,
      "grad_norm": 0.15144968032836914,
      "learning_rate": 9.984113417406939e-05,
      "loss": 1.4128,
      "mean_token_accuracy": 0.7451657593250275,
      "num_tokens": 293179588.0,
      "step": 4080
    },
    {
      "entropy": 0.9735559523105621,
      "epoch": 1.1147451621695286,
      "grad_norm": 0.2099827527999878,
      "learning_rate": 9.98387320845517e-05,
      "loss": 1.4082,
      "mean_token_accuracy": 0.7413778752088547,
      "num_tokens": 293902990.0,
      "step": 4090
    },
    {
      "entropy": 0.9711895674467087,
      "epoch": 1.1174707004633415,
      "grad_norm": 0.30111899971961975,
      "learning_rate": 9.983631200028808e-05,
      "loss": 1.4069,
      "mean_token_accuracy": 0.7399752870202064,
      "num_tokens": 294628257.0,
      "step": 4100
    },
    {
      "entropy": 0.967848864197731,
      "epoch": 1.1201962387571545,
      "grad_norm": 0.14860115945339203,
      "learning_rate": 9.983387392215238e-05,
      "loss": 1.4035,
      "mean_token_accuracy": 0.7399855732917786,
      "num_tokens": 295343589.0,
      "step": 4110
    },
    {
      "entropy": 0.9747548431158066,
      "epoch": 1.1229217770509676,
      "grad_norm": 0.15380148589611053,
      "learning_rate": 9.983141785102483e-05,
      "loss": 1.4168,
      "mean_token_accuracy": 0.7410910993814468,
      "num_tokens": 296066401.0,
      "step": 4120
    },
    {
      "entropy": 0.9697711855173111,
      "epoch": 1.1256473153447806,
      "grad_norm": 0.12143965810537338,
      "learning_rate": 9.982894378779226e-05,
      "loss": 1.4141,
      "mean_token_accuracy": 0.7393874973058701,
      "num_tokens": 296784664.0,
      "step": 4130
    },
    {
      "entropy": 0.970097741484642,
      "epoch": 1.1283728536385937,
      "grad_norm": 0.11673377454280853,
      "learning_rate": 9.982645173334794e-05,
      "loss": 1.4089,
      "mean_token_accuracy": 0.7406289339065552,
      "num_tokens": 297499538.0,
      "step": 4140
    },
    {
      "entropy": 0.9692523866891861,
      "epoch": 1.1310983919324067,
      "grad_norm": 0.1403290182352066,
      "learning_rate": 9.982394168859164e-05,
      "loss": 1.4131,
      "mean_token_accuracy": 0.7348797231912613,
      "num_tokens": 298199689.0,
      "step": 4150
    },
    {
      "entropy": 0.9668681174516678,
      "epoch": 1.1338239302262196,
      "grad_norm": 0.1316184103488922,
      "learning_rate": 9.982141365442966e-05,
      "loss": 1.4069,
      "mean_token_accuracy": 0.7375933572649955,
      "num_tokens": 298910195.0,
      "step": 4160
    },
    {
      "entropy": 0.9702906340360642,
      "epoch": 1.1365494685200328,
      "grad_norm": 0.12228190153837204,
      "learning_rate": 9.981886763177476e-05,
      "loss": 1.4151,
      "mean_token_accuracy": 0.7387032642960548,
      "num_tokens": 299620451.0,
      "step": 4170
    },
    {
      "entropy": 0.9716745302081108,
      "epoch": 1.1392750068138457,
      "grad_norm": 0.1245054379105568,
      "learning_rate": 9.98163036215462e-05,
      "loss": 1.4065,
      "mean_token_accuracy": 0.7411650583148003,
      "num_tokens": 300344107.0,
      "step": 4180
    },
    {
      "entropy": 0.9720513239502907,
      "epoch": 1.1420005451076587,
      "grad_norm": 0.15862545371055603,
      "learning_rate": 9.981372162466973e-05,
      "loss": 1.4124,
      "mean_token_accuracy": 0.7386313796043396,
      "num_tokens": 301061829.0,
      "step": 4190
    },
    {
      "entropy": 0.9701609164476395,
      "epoch": 1.1447260834014719,
      "grad_norm": 0.18401804566383362,
      "learning_rate": 9.981112164207763e-05,
      "loss": 1.4042,
      "mean_token_accuracy": 0.7445456832647324,
      "num_tokens": 301792603.0,
      "step": 4200
    },
    {
      "entropy": 0.9677683159708976,
      "epoch": 1.1474516216952848,
      "grad_norm": 0.12356293946504593,
      "learning_rate": 9.980850367470863e-05,
      "loss": 1.399,
      "mean_token_accuracy": 0.7404949709773063,
      "num_tokens": 302512106.0,
      "step": 4210
    },
    {
      "entropy": 0.9647759571671486,
      "epoch": 1.1501771599890978,
      "grad_norm": 0.13986940681934357,
      "learning_rate": 9.980586772350798e-05,
      "loss": 1.3967,
      "mean_token_accuracy": 0.7414400741457939,
      "num_tokens": 303239734.0,
      "step": 4220
    },
    {
      "entropy": 0.9754222020506859,
      "epoch": 1.152902698282911,
      "grad_norm": 0.1358058601617813,
      "learning_rate": 9.980321378942741e-05,
      "loss": 1.4137,
      "mean_token_accuracy": 0.7352504953742027,
      "num_tokens": 303949407.0,
      "step": 4230
    },
    {
      "entropy": 0.9723289042711258,
      "epoch": 1.1556282365767239,
      "grad_norm": 0.1265636384487152,
      "learning_rate": 9.980054187342516e-05,
      "loss": 1.4089,
      "mean_token_accuracy": 0.7384456306695938,
      "num_tokens": 304662945.0,
      "step": 4240
    },
    {
      "entropy": 0.974214768409729,
      "epoch": 1.158353774870537,
      "grad_norm": 0.12341339141130447,
      "learning_rate": 9.979785197646593e-05,
      "loss": 1.413,
      "mean_token_accuracy": 0.7399124011397362,
      "num_tokens": 305380069.0,
      "step": 4250
    },
    {
      "entropy": 0.9702006042003631,
      "epoch": 1.16107931316435,
      "grad_norm": 0.1514752358198166,
      "learning_rate": 9.979514409952096e-05,
      "loss": 1.412,
      "mean_token_accuracy": 0.7410535365343094,
      "num_tokens": 306097964.0,
      "step": 4260
    },
    {
      "entropy": 0.9635045930743218,
      "epoch": 1.163804851458163,
      "grad_norm": 0.09730468690395355,
      "learning_rate": 9.979241824356791e-05,
      "loss": 1.4014,
      "mean_token_accuracy": 0.7407137379050255,
      "num_tokens": 306810361.0,
      "step": 4270
    },
    {
      "entropy": 0.9600163742899894,
      "epoch": 1.166530389751976,
      "grad_norm": 0.19360746443271637,
      "learning_rate": 9.978967440959101e-05,
      "loss": 1.3968,
      "mean_token_accuracy": 0.7477250754833221,
      "num_tokens": 307552298.0,
      "step": 4280
    },
    {
      "entropy": 0.9812612414360047,
      "epoch": 1.169255928045789,
      "grad_norm": 0.09811007231473923,
      "learning_rate": 9.978691259858097e-05,
      "loss": 1.4168,
      "mean_token_accuracy": 0.7393087267875671,
      "num_tokens": 308269441.0,
      "step": 4290
    },
    {
      "entropy": 0.9744583383202553,
      "epoch": 1.171981466339602,
      "grad_norm": 0.1087038516998291,
      "learning_rate": 9.978413281153491e-05,
      "loss": 1.4132,
      "mean_token_accuracy": 0.7385753050446511,
      "num_tokens": 308987989.0,
      "step": 4300
    },
    {
      "entropy": 0.9722009614109993,
      "epoch": 1.1747070046334152,
      "grad_norm": 0.25121307373046875,
      "learning_rate": 9.978133504945655e-05,
      "loss": 1.406,
      "mean_token_accuracy": 0.7378881454467774,
      "num_tokens": 309705847.0,
      "step": 4310
    },
    {
      "entropy": 0.9756908550858497,
      "epoch": 1.177432542927228,
      "grad_norm": 0.14021795988082886,
      "learning_rate": 9.977851931335602e-05,
      "loss": 1.4008,
      "mean_token_accuracy": 0.7398674771189689,
      "num_tokens": 310432150.0,
      "step": 4320
    },
    {
      "entropy": 0.9696485593914985,
      "epoch": 1.180158081221041,
      "grad_norm": 0.13773490488529205,
      "learning_rate": 9.977568560424997e-05,
      "loss": 1.4141,
      "mean_token_accuracy": 0.7451538845896721,
      "num_tokens": 311161749.0,
      "step": 4330
    },
    {
      "entropy": 0.9696112379431725,
      "epoch": 1.1828836195148542,
      "grad_norm": 0.11383041739463806,
      "learning_rate": 9.977283392316154e-05,
      "loss": 1.4009,
      "mean_token_accuracy": 0.741061507165432,
      "num_tokens": 311887750.0,
      "step": 4340
    },
    {
      "entropy": 0.9739535748958588,
      "epoch": 1.1856091578086672,
      "grad_norm": 0.14352960884571075,
      "learning_rate": 9.976996427112036e-05,
      "loss": 1.4172,
      "mean_token_accuracy": 0.7372234463691711,
      "num_tokens": 312602425.0,
      "step": 4350
    },
    {
      "entropy": 0.9777967765927315,
      "epoch": 1.1883346961024803,
      "grad_norm": 0.14844806492328644,
      "learning_rate": 9.976707664916254e-05,
      "loss": 1.4069,
      "mean_token_accuracy": 0.7373544290661812,
      "num_tokens": 313332944.0,
      "step": 4360
    },
    {
      "entropy": 0.973545515537262,
      "epoch": 1.1910602343962933,
      "grad_norm": 0.1698267161846161,
      "learning_rate": 9.97641710583307e-05,
      "loss": 1.4068,
      "mean_token_accuracy": 0.7357580438256264,
      "num_tokens": 314046985.0,
      "step": 4370
    },
    {
      "entropy": 0.9729410991072655,
      "epoch": 1.1937857726901062,
      "grad_norm": 0.21367119252681732,
      "learning_rate": 9.976124749967391e-05,
      "loss": 1.4061,
      "mean_token_accuracy": 0.7449333608150482,
      "num_tokens": 314784201.0,
      "step": 4380
    },
    {
      "entropy": 0.9695119127631188,
      "epoch": 1.1965113109839194,
      "grad_norm": 0.1914973258972168,
      "learning_rate": 9.975830597424777e-05,
      "loss": 1.398,
      "mean_token_accuracy": 0.7413497671484948,
      "num_tokens": 315515584.0,
      "step": 4390
    },
    {
      "entropy": 0.9745991230010986,
      "epoch": 1.1992368492777323,
      "grad_norm": 0.09760070592164993,
      "learning_rate": 9.97553464831143e-05,
      "loss": 1.4097,
      "mean_token_accuracy": 0.7376661255955697,
      "num_tokens": 316231773.0,
      "step": 4400
    },
    {
      "entropy": 0.9677482321858406,
      "epoch": 1.2019623875715455,
      "grad_norm": 0.1391540765762329,
      "learning_rate": 9.975236902734213e-05,
      "loss": 1.4046,
      "mean_token_accuracy": 0.7410488307476044,
      "num_tokens": 316959973.0,
      "step": 4410
    },
    {
      "entropy": 0.9732399925589561,
      "epoch": 1.2046879258653584,
      "grad_norm": 0.11403980106115341,
      "learning_rate": 9.974937360800622e-05,
      "loss": 1.4102,
      "mean_token_accuracy": 0.736595906317234,
      "num_tokens": 317668555.0,
      "step": 4420
    },
    {
      "entropy": 0.9660505890846253,
      "epoch": 1.2074134641591714,
      "grad_norm": 0.09890380501747131,
      "learning_rate": 9.974636022618813e-05,
      "loss": 1.4028,
      "mean_token_accuracy": 0.7411175265908241,
      "num_tokens": 318391002.0,
      "step": 4430
    },
    {
      "entropy": 0.9746992096304894,
      "epoch": 1.2101390024529846,
      "grad_norm": 0.1454600989818573,
      "learning_rate": 9.974332888297589e-05,
      "loss": 1.4015,
      "mean_token_accuracy": 0.742490966618061,
      "num_tokens": 319121319.0,
      "step": 4440
    },
    {
      "entropy": 0.9703999727964401,
      "epoch": 1.2128645407467975,
      "grad_norm": 0.14531758427619934,
      "learning_rate": 9.974027957946394e-05,
      "loss": 1.4077,
      "mean_token_accuracy": 0.7393903106451034,
      "num_tokens": 319844030.0,
      "step": 4450
    },
    {
      "entropy": 0.9730675384402275,
      "epoch": 1.2155900790406104,
      "grad_norm": 0.09914612770080566,
      "learning_rate": 9.973721231675331e-05,
      "loss": 1.4034,
      "mean_token_accuracy": 0.7403646603226661,
      "num_tokens": 320571272.0,
      "step": 4460
    },
    {
      "entropy": 0.9742552965879441,
      "epoch": 1.2183156173344236,
      "grad_norm": 0.1258276253938675,
      "learning_rate": 9.973412709595144e-05,
      "loss": 1.4077,
      "mean_token_accuracy": 0.7421027988195419,
      "num_tokens": 321297398.0,
      "step": 4470
    },
    {
      "entropy": 0.9606939122080803,
      "epoch": 1.2210411556282366,
      "grad_norm": 0.11029420793056488,
      "learning_rate": 9.97310239181723e-05,
      "loss": 1.3993,
      "mean_token_accuracy": 0.7463144496083259,
      "num_tokens": 322035012.0,
      "step": 4480
    },
    {
      "entropy": 0.9743346512317658,
      "epoch": 1.2237666939220495,
      "grad_norm": 0.13282760977745056,
      "learning_rate": 9.97279027845363e-05,
      "loss": 1.4117,
      "mean_token_accuracy": 0.7394008815288544,
      "num_tokens": 322759503.0,
      "step": 4490
    },
    {
      "entropy": 0.9749991357326507,
      "epoch": 1.2264922322158627,
      "grad_norm": 0.30340775847435,
      "learning_rate": 9.972476369617036e-05,
      "loss": 1.4215,
      "mean_token_accuracy": 0.7399731442332268,
      "num_tokens": 323471760.0,
      "step": 4500
    },
    {
      "entropy": 0.9715635746717453,
      "epoch": 1.2292177705096756,
      "grad_norm": 0.19635392725467682,
      "learning_rate": 9.972160665420788e-05,
      "loss": 1.411,
      "mean_token_accuracy": 0.738856740295887,
      "num_tokens": 324191070.0,
      "step": 4510
    },
    {
      "entropy": 0.9710773929953576,
      "epoch": 1.2319433088034888,
      "grad_norm": 0.16573446989059448,
      "learning_rate": 9.971843165978873e-05,
      "loss": 1.3998,
      "mean_token_accuracy": 0.7402691707015038,
      "num_tokens": 324915048.0,
      "step": 4520
    },
    {
      "entropy": 0.9762237891554832,
      "epoch": 1.2346688470973017,
      "grad_norm": 0.1468210071325302,
      "learning_rate": 9.97152387140593e-05,
      "loss": 1.4127,
      "mean_token_accuracy": 0.7367881044745446,
      "num_tokens": 325626112.0,
      "step": 4530
    },
    {
      "entropy": 0.9673715129494667,
      "epoch": 1.2373943853911147,
      "grad_norm": 0.15306654572486877,
      "learning_rate": 9.97120278181724e-05,
      "loss": 1.4068,
      "mean_token_accuracy": 0.7392708584666252,
      "num_tokens": 326349540.0,
      "step": 4540
    },
    {
      "entropy": 0.9674241647124291,
      "epoch": 1.2401199236849278,
      "grad_norm": 0.14154525101184845,
      "learning_rate": 9.970879897328738e-05,
      "loss": 1.4022,
      "mean_token_accuracy": 0.739752496778965,
      "num_tokens": 327070191.0,
      "step": 4550
    },
    {
      "entropy": 0.9727215856313706,
      "epoch": 1.2428454619787408,
      "grad_norm": 0.11445024609565735,
      "learning_rate": 9.970555218057002e-05,
      "loss": 1.4128,
      "mean_token_accuracy": 0.7393271207809449,
      "num_tokens": 327785310.0,
      "step": 4560
    },
    {
      "entropy": 0.9723274782299995,
      "epoch": 1.2455710002725537,
      "grad_norm": 0.138005793094635,
      "learning_rate": 9.970228744119264e-05,
      "loss": 1.4137,
      "mean_token_accuracy": 0.7351332068443298,
      "num_tokens": 328491172.0,
      "step": 4570
    },
    {
      "entropy": 0.9660980984568596,
      "epoch": 1.248296538566367,
      "grad_norm": 0.13529391586780548,
      "learning_rate": 9.969900475633395e-05,
      "loss": 1.4026,
      "mean_token_accuracy": 0.7479706987738609,
      "num_tokens": 329232744.0,
      "step": 4580
    },
    {
      "entropy": 0.9704912841320038,
      "epoch": 1.2510220768601799,
      "grad_norm": 0.162317156791687,
      "learning_rate": 9.969570412717925e-05,
      "loss": 1.4057,
      "mean_token_accuracy": 0.7417619049549102,
      "num_tokens": 329966706.0,
      "step": 4590
    },
    {
      "entropy": 0.9835660576820373,
      "epoch": 1.2537476151539928,
      "grad_norm": 0.11597214639186859,
      "learning_rate": 9.969238555492025e-05,
      "loss": 1.4146,
      "mean_token_accuracy": 0.7364133894443512,
      "num_tokens": 330682714.0,
      "step": 4600
    },
    {
      "entropy": 0.9688590139150619,
      "epoch": 1.256473153447806,
      "grad_norm": 0.12225977331399918,
      "learning_rate": 9.968904904075511e-05,
      "loss": 1.4072,
      "mean_token_accuracy": 0.7402441427111626,
      "num_tokens": 331401822.0,
      "step": 4610
    },
    {
      "entropy": 0.9649090647697449,
      "epoch": 1.259198691741619,
      "grad_norm": 0.10383719950914383,
      "learning_rate": 9.968569458588857e-05,
      "loss": 1.3949,
      "mean_token_accuracy": 0.7428733959794045,
      "num_tokens": 332125457.0,
      "step": 4620
    },
    {
      "entropy": 0.9636820971965789,
      "epoch": 1.261924230035432,
      "grad_norm": 0.2340947836637497,
      "learning_rate": 9.968232219153175e-05,
      "loss": 1.3969,
      "mean_token_accuracy": 0.7443041607737542,
      "num_tokens": 332857765.0,
      "step": 4630
    },
    {
      "entropy": 0.9748477324843406,
      "epoch": 1.264649768329245,
      "grad_norm": 0.1700190305709839,
      "learning_rate": 9.967893185890232e-05,
      "loss": 1.4128,
      "mean_token_accuracy": 0.7405909344553947,
      "num_tokens": 333578614.0,
      "step": 4640
    },
    {
      "entropy": 0.9795809105038643,
      "epoch": 1.267375306623058,
      "grad_norm": 0.1715494692325592,
      "learning_rate": 9.967552358922434e-05,
      "loss": 1.4232,
      "mean_token_accuracy": 0.7324912399053574,
      "num_tokens": 334268107.0,
      "step": 4650
    },
    {
      "entropy": 0.9613471925258636,
      "epoch": 1.2701008449168711,
      "grad_norm": 0.11859522014856339,
      "learning_rate": 9.967209738372845e-05,
      "loss": 1.4037,
      "mean_token_accuracy": 0.7405437037348748,
      "num_tokens": 334974099.0,
      "step": 4660
    },
    {
      "entropy": 0.9693601042032242,
      "epoch": 1.272826383210684,
      "grad_norm": 0.12638618052005768,
      "learning_rate": 9.966865324365167e-05,
      "loss": 1.3971,
      "mean_token_accuracy": 0.7439545139670372,
      "num_tokens": 335712163.0,
      "step": 4670
    },
    {
      "entropy": 0.9626713693141937,
      "epoch": 1.2755519215044973,
      "grad_norm": 0.13341055810451508,
      "learning_rate": 9.966519117023756e-05,
      "loss": 1.4017,
      "mean_token_accuracy": 0.7433236762881279,
      "num_tokens": 336432921.0,
      "step": 4680
    },
    {
      "entropy": 0.9728736892342568,
      "epoch": 1.2782774597983102,
      "grad_norm": 0.14286339282989502,
      "learning_rate": 9.966171116473614e-05,
      "loss": 1.404,
      "mean_token_accuracy": 0.7400946334004402,
      "num_tokens": 337153765.0,
      "step": 4690
    },
    {
      "entropy": 0.9666111811995506,
      "epoch": 1.2810029980921231,
      "grad_norm": 0.14786115288734436,
      "learning_rate": 9.96582132284039e-05,
      "loss": 1.3987,
      "mean_token_accuracy": 0.7426782041788101,
      "num_tokens": 337881884.0,
      "step": 4700
    },
    {
      "entropy": 0.9663512572646141,
      "epoch": 1.283728536385936,
      "grad_norm": 0.11639788001775742,
      "learning_rate": 9.965469736250377e-05,
      "loss": 1.4023,
      "mean_token_accuracy": 0.7427395194768905,
      "num_tokens": 338610104.0,
      "step": 4710
    },
    {
      "entropy": 0.9779267340898514,
      "epoch": 1.2864540746797493,
      "grad_norm": 0.09598447382450104,
      "learning_rate": 9.965116356830523e-05,
      "loss": 1.4148,
      "mean_token_accuracy": 0.7342917695641518,
      "num_tokens": 339318927.0,
      "step": 4720
    },
    {
      "entropy": 0.9727552771568299,
      "epoch": 1.2891796129735622,
      "grad_norm": 0.10459399223327637,
      "learning_rate": 9.964761184708418e-05,
      "loss": 1.4069,
      "mean_token_accuracy": 0.7402333587408065,
      "num_tokens": 340037954.0,
      "step": 4730
    },
    {
      "entropy": 0.970164880156517,
      "epoch": 1.2919051512673754,
      "grad_norm": 0.14718195796012878,
      "learning_rate": 9.964404220012298e-05,
      "loss": 1.4051,
      "mean_token_accuracy": 0.7438466176390648,
      "num_tokens": 340769677.0,
      "step": 4740
    },
    {
      "entropy": 0.9741303488612175,
      "epoch": 1.2946306895611883,
      "grad_norm": 0.10659249126911163,
      "learning_rate": 9.96404546287105e-05,
      "loss": 1.4108,
      "mean_token_accuracy": 0.7378978326916694,
      "num_tokens": 341482922.0,
      "step": 4750
    },
    {
      "entropy": 0.964579838514328,
      "epoch": 1.2973562278550013,
      "grad_norm": 0.10390475392341614,
      "learning_rate": 9.963684913414208e-05,
      "loss": 1.3923,
      "mean_token_accuracy": 0.7442897364497185,
      "num_tokens": 342224002.0,
      "step": 4760
    },
    {
      "entropy": 0.9747483864426613,
      "epoch": 1.3000817661488144,
      "grad_norm": 0.24640607833862305,
      "learning_rate": 9.963322571771949e-05,
      "loss": 1.4091,
      "mean_token_accuracy": 0.7386063888669014,
      "num_tokens": 342936013.0,
      "step": 4770
    },
    {
      "entropy": 0.9741189420223236,
      "epoch": 1.3028073044426274,
      "grad_norm": 0.11026379466056824,
      "learning_rate": 9.9629584380751e-05,
      "loss": 1.4077,
      "mean_token_accuracy": 0.7361203297972679,
      "num_tokens": 343651597.0,
      "step": 4780
    },
    {
      "entropy": 0.9651884615421296,
      "epoch": 1.3055328427364405,
      "grad_norm": 0.18231336772441864,
      "learning_rate": 9.962592512455138e-05,
      "loss": 1.4015,
      "mean_token_accuracy": 0.7450736105442047,
      "num_tokens": 344387553.0,
      "step": 4790
    },
    {
      "entropy": 0.961803525686264,
      "epoch": 1.3082583810302535,
      "grad_norm": 0.10563954710960388,
      "learning_rate": 9.962224795044184e-05,
      "loss": 1.3941,
      "mean_token_accuracy": 0.7397433206439018,
      "num_tokens": 345102338.0,
      "step": 4800
    },
    {
      "entropy": 0.9752078264951706,
      "epoch": 1.3109839193240664,
      "grad_norm": 0.08828509598970413,
      "learning_rate": 9.961855285975001e-05,
      "loss": 1.4062,
      "mean_token_accuracy": 0.7382546260952949,
      "num_tokens": 345815930.0,
      "step": 4810
    },
    {
      "entropy": 0.974830986559391,
      "epoch": 1.3137094576178796,
      "grad_norm": 0.14658141136169434,
      "learning_rate": 9.961483985381009e-05,
      "loss": 1.4154,
      "mean_token_accuracy": 0.7359081193804741,
      "num_tokens": 346522652.0,
      "step": 4820
    },
    {
      "entropy": 0.9659509539604187,
      "epoch": 1.3164349959116926,
      "grad_norm": 0.11177151650190353,
      "learning_rate": 9.961110893396263e-05,
      "loss": 1.3946,
      "mean_token_accuracy": 0.7395478710532188,
      "num_tokens": 347233138.0,
      "step": 4830
    },
    {
      "entropy": 0.9693713441491127,
      "epoch": 1.3191605342055057,
      "grad_norm": 0.158463716506958,
      "learning_rate": 9.96073601015548e-05,
      "loss": 1.4044,
      "mean_token_accuracy": 0.7397462442517281,
      "num_tokens": 347951579.0,
      "step": 4840
    },
    {
      "entropy": 0.9697089105844497,
      "epoch": 1.3218860724993187,
      "grad_norm": 0.1285538375377655,
      "learning_rate": 9.96035933579401e-05,
      "loss": 1.402,
      "mean_token_accuracy": 0.7475307866930961,
      "num_tokens": 348693113.0,
      "step": 4850
    },
    {
      "entropy": 0.9771092981100082,
      "epoch": 1.3246116107931316,
      "grad_norm": 0.1073615625500679,
      "learning_rate": 9.959980870447854e-05,
      "loss": 1.4112,
      "mean_token_accuracy": 0.7395440816879273,
      "num_tokens": 349420285.0,
      "step": 4860
    },
    {
      "entropy": 0.9649139031767845,
      "epoch": 1.3273371490869446,
      "grad_norm": 0.11734608560800552,
      "learning_rate": 9.959600614253663e-05,
      "loss": 1.409,
      "mean_token_accuracy": 0.7405727341771126,
      "num_tokens": 350135291.0,
      "step": 4870
    },
    {
      "entropy": 0.9652960941195488,
      "epoch": 1.3300626873807577,
      "grad_norm": 0.10706183314323425,
      "learning_rate": 9.959218567348732e-05,
      "loss": 1.3983,
      "mean_token_accuracy": 0.7411574274301529,
      "num_tokens": 350851336.0,
      "step": 4880
    },
    {
      "entropy": 0.9691098436713219,
      "epoch": 1.3327882256745707,
      "grad_norm": 0.12964962422847748,
      "learning_rate": 9.958834729871002e-05,
      "loss": 1.4151,
      "mean_token_accuracy": 0.7378066569566727,
      "num_tokens": 351555913.0,
      "step": 4890
    },
    {
      "entropy": 0.9715841427445412,
      "epoch": 1.3355137639683838,
      "grad_norm": 0.10759638249874115,
      "learning_rate": 9.95844910195906e-05,
      "loss": 1.4039,
      "mean_token_accuracy": 0.7350040569901466,
      "num_tokens": 352269755.0,
      "step": 4900
    },
    {
      "entropy": 0.974685150384903,
      "epoch": 1.3382393022621968,
      "grad_norm": 0.1167893037199974,
      "learning_rate": 9.958061683752143e-05,
      "loss": 1.4088,
      "mean_token_accuracy": 0.7382029041647911,
      "num_tokens": 352996391.0,
      "step": 4910
    },
    {
      "entropy": 0.9621032327413559,
      "epoch": 1.3409648405560097,
      "grad_norm": 0.10185514390468597,
      "learning_rate": 9.957672475390129e-05,
      "loss": 1.3953,
      "mean_token_accuracy": 0.7431463569402694,
      "num_tokens": 353717924.0,
      "step": 4920
    },
    {
      "entropy": 0.9662080585956574,
      "epoch": 1.343690378849823,
      "grad_norm": 0.11236744374036789,
      "learning_rate": 9.957281477013548e-05,
      "loss": 1.4055,
      "mean_token_accuracy": 0.7385106563568116,
      "num_tokens": 354429863.0,
      "step": 4930
    },
    {
      "entropy": 0.9646960616111755,
      "epoch": 1.3464159171436358,
      "grad_norm": 0.10907775908708572,
      "learning_rate": 9.956888688763573e-05,
      "loss": 1.3959,
      "mean_token_accuracy": 0.7475621044635773,
      "num_tokens": 355175867.0,
      "step": 4940
    },
    {
      "entropy": 0.9851981371641159,
      "epoch": 1.349141455437449,
      "grad_norm": 0.23139843344688416,
      "learning_rate": 9.956494110782022e-05,
      "loss": 1.4173,
      "mean_token_accuracy": 0.7367147475481033,
      "num_tokens": 355898678.0,
      "step": 4950
    },
    {
      "entropy": 0.972486612200737,
      "epoch": 1.351866993731262,
      "grad_norm": 0.10554983466863632,
      "learning_rate": 9.956097743211363e-05,
      "loss": 1.4046,
      "mean_token_accuracy": 0.7401088133454323,
      "num_tokens": 356631902.0,
      "step": 4960
    },
    {
      "entropy": 0.9727440178394318,
      "epoch": 1.354592532025075,
      "grad_norm": 0.15384016931056976,
      "learning_rate": 9.95569958619471e-05,
      "loss": 1.4044,
      "mean_token_accuracy": 0.7396053314208985,
      "num_tokens": 357348213.0,
      "step": 4970
    },
    {
      "entropy": 0.9753688693046569,
      "epoch": 1.3573180703188878,
      "grad_norm": 0.11177823692560196,
      "learning_rate": 9.955299639875818e-05,
      "loss": 1.4169,
      "mean_token_accuracy": 0.736075484752655,
      "num_tokens": 358047933.0,
      "step": 4980
    },
    {
      "entropy": 0.9673207774758339,
      "epoch": 1.360043608612701,
      "grad_norm": 0.09317687153816223,
      "learning_rate": 9.954897904399093e-05,
      "loss": 1.4044,
      "mean_token_accuracy": 0.7369695752859116,
      "num_tokens": 358762622.0,
      "step": 4990
    },
    {
      "entropy": 0.9734054937958717,
      "epoch": 1.362769146906514,
      "grad_norm": 0.17638979852199554,
      "learning_rate": 9.954494379909585e-05,
      "loss": 1.4028,
      "mean_token_accuracy": 0.7384358510375023,
      "num_tokens": 359480737.0,
      "step": 5000
    },
    {
      "entropy": 0.97704269438982,
      "epoch": 1.3654946852003271,
      "grad_norm": 0.11890945583581924,
      "learning_rate": 9.954089066552991e-05,
      "loss": 1.4066,
      "mean_token_accuracy": 0.7409640535712242,
      "num_tokens": 360205405.0,
      "step": 5010
    },
    {
      "entropy": 0.9697741746902466,
      "epoch": 1.36822022349414,
      "grad_norm": 0.21392837166786194,
      "learning_rate": 9.953681964475652e-05,
      "loss": 1.4084,
      "mean_token_accuracy": 0.737616577744484,
      "num_tokens": 360916851.0,
      "step": 5020
    },
    {
      "entropy": 0.9659984797239304,
      "epoch": 1.370945761787953,
      "grad_norm": 0.10474559664726257,
      "learning_rate": 9.953273073824557e-05,
      "loss": 1.3971,
      "mean_token_accuracy": 0.74130250364542,
      "num_tokens": 361639152.0,
      "step": 5030
    },
    {
      "entropy": 0.9709238648414612,
      "epoch": 1.3736713000817662,
      "grad_norm": 0.1659403145313263,
      "learning_rate": 9.95286239474734e-05,
      "loss": 1.4007,
      "mean_token_accuracy": 0.7374826833605767,
      "num_tokens": 362347888.0,
      "step": 5040
    },
    {
      "entropy": 0.9627052456140518,
      "epoch": 1.3763968383755791,
      "grad_norm": 0.12752042710781097,
      "learning_rate": 9.95244992739228e-05,
      "loss": 1.4015,
      "mean_token_accuracy": 0.7374019876122475,
      "num_tokens": 363055074.0,
      "step": 5050
    },
    {
      "entropy": 0.9730372503399849,
      "epoch": 1.3791223766693923,
      "grad_norm": 0.10037846863269806,
      "learning_rate": 9.952035671908302e-05,
      "loss": 1.4055,
      "mean_token_accuracy": 0.7396304190158844,
      "num_tokens": 363770036.0,
      "step": 5060
    },
    {
      "entropy": 0.9712066382169724,
      "epoch": 1.3818479149632052,
      "grad_norm": 0.13504239916801453,
      "learning_rate": 9.951619628444977e-05,
      "loss": 1.4064,
      "mean_token_accuracy": 0.7389048680663108,
      "num_tokens": 364480178.0,
      "step": 5070
    },
    {
      "entropy": 0.9749246120452881,
      "epoch": 1.3845734532570182,
      "grad_norm": 0.13781282305717468,
      "learning_rate": 9.951201797152523e-05,
      "loss": 1.4065,
      "mean_token_accuracy": 0.7400273680686951,
      "num_tokens": 365187178.0,
      "step": 5080
    },
    {
      "entropy": 0.971899363398552,
      "epoch": 1.3872989915508314,
      "grad_norm": 0.1477481722831726,
      "learning_rate": 9.9507821781818e-05,
      "loss": 1.4158,
      "mean_token_accuracy": 0.730982817709446,
      "num_tokens": 365883898.0,
      "step": 5090
    },
    {
      "entropy": 0.966162946820259,
      "epoch": 1.3900245298446443,
      "grad_norm": 0.1793452352285385,
      "learning_rate": 9.950360771684316e-05,
      "loss": 1.403,
      "mean_token_accuracy": 0.7393653288483619,
      "num_tokens": 366602829.0,
      "step": 5100
    },
    {
      "entropy": 0.9741988733410836,
      "epoch": 1.3927500681384575,
      "grad_norm": 0.1578950583934784,
      "learning_rate": 9.949937577812223e-05,
      "loss": 1.4071,
      "mean_token_accuracy": 0.7414002895355225,
      "num_tokens": 367316769.0,
      "step": 5110
    },
    {
      "entropy": 0.9581542372703552,
      "epoch": 1.3954756064322704,
      "grad_norm": 0.1000090092420578,
      "learning_rate": 9.949512596718323e-05,
      "loss": 1.3946,
      "mean_token_accuracy": 0.7404720976948738,
      "num_tokens": 368032771.0,
      "step": 5120
    },
    {
      "entropy": 0.9679626032710076,
      "epoch": 1.3982011447260834,
      "grad_norm": 0.108095683157444,
      "learning_rate": 9.949085828556053e-05,
      "loss": 1.4133,
      "mean_token_accuracy": 0.735573785007,
      "num_tokens": 368734920.0,
      "step": 5130
    },
    {
      "entropy": 0.9785879150032997,
      "epoch": 1.4009266830198963,
      "grad_norm": 0.1070920005440712,
      "learning_rate": 9.948657273479507e-05,
      "loss": 1.4184,
      "mean_token_accuracy": 0.73294368237257,
      "num_tokens": 369438117.0,
      "step": 5140
    },
    {
      "entropy": 0.9624388098716736,
      "epoch": 1.4036522213137095,
      "grad_norm": 0.11080455034971237,
      "learning_rate": 9.948226931643417e-05,
      "loss": 1.3899,
      "mean_token_accuracy": 0.7419885486364365,
      "num_tokens": 370169171.0,
      "step": 5150
    },
    {
      "entropy": 0.9686688229441642,
      "epoch": 1.4063777596075224,
      "grad_norm": 0.11958616226911545,
      "learning_rate": 9.94779480320316e-05,
      "loss": 1.402,
      "mean_token_accuracy": 0.7391363710165024,
      "num_tokens": 370885893.0,
      "step": 5160
    },
    {
      "entropy": 0.9716649666428566,
      "epoch": 1.4091032979013356,
      "grad_norm": 0.26983004808425903,
      "learning_rate": 9.947360888314765e-05,
      "loss": 1.4,
      "mean_token_accuracy": 0.7359407395124435,
      "num_tokens": 371592228.0,
      "step": 5170
    },
    {
      "entropy": 0.9631922513246536,
      "epoch": 1.4118288361951485,
      "grad_norm": 0.08574482053518295,
      "learning_rate": 9.946925187134897e-05,
      "loss": 1.4038,
      "mean_token_accuracy": 0.7413485586643219,
      "num_tokens": 372304040.0,
      "step": 5180
    },
    {
      "entropy": 0.9777448117733002,
      "epoch": 1.4145543744889615,
      "grad_norm": 0.12568679451942444,
      "learning_rate": 9.94648769982087e-05,
      "loss": 1.4055,
      "mean_token_accuracy": 0.7402204349637032,
      "num_tokens": 373032183.0,
      "step": 5190
    },
    {
      "entropy": 0.9667919337749481,
      "epoch": 1.4172799127827747,
      "grad_norm": 0.10519536584615707,
      "learning_rate": 9.946048426530646e-05,
      "loss": 1.4047,
      "mean_token_accuracy": 0.7366608127951622,
      "num_tokens": 373740579.0,
      "step": 5200
    },
    {
      "entropy": 0.9548894897103309,
      "epoch": 1.4200054510765876,
      "grad_norm": 0.13211771845817566,
      "learning_rate": 9.945607367422825e-05,
      "loss": 1.4001,
      "mean_token_accuracy": 0.7390782698988915,
      "num_tokens": 374439723.0,
      "step": 5210
    },
    {
      "entropy": 0.9750136703252792,
      "epoch": 1.4227309893704008,
      "grad_norm": 0.12971355020999908,
      "learning_rate": 9.945164522656659e-05,
      "loss": 1.4114,
      "mean_token_accuracy": 0.7375137954950333,
      "num_tokens": 375151408.0,
      "step": 5220
    },
    {
      "entropy": 0.9739547118544578,
      "epoch": 1.4254565276642137,
      "grad_norm": 0.14966237545013428,
      "learning_rate": 9.944719892392038e-05,
      "loss": 1.4142,
      "mean_token_accuracy": 0.7377022936940193,
      "num_tokens": 375870416.0,
      "step": 5230
    },
    {
      "entropy": 0.9648297607898713,
      "epoch": 1.4281820659580267,
      "grad_norm": 0.36714956164360046,
      "learning_rate": 9.944273476789503e-05,
      "loss": 1.3945,
      "mean_token_accuracy": 0.7440789759159088,
      "num_tokens": 376599833.0,
      "step": 5240
    },
    {
      "entropy": 0.9652238085865974,
      "epoch": 1.4309076042518396,
      "grad_norm": 0.1441311538219452,
      "learning_rate": 9.943825276010235e-05,
      "loss": 1.4013,
      "mean_token_accuracy": 0.740245969593525,
      "num_tokens": 377316609.0,
      "step": 5250
    },
    {
      "entropy": 0.9736945912241936,
      "epoch": 1.4336331425456528,
      "grad_norm": 0.10831708461046219,
      "learning_rate": 9.943375290216061e-05,
      "loss": 1.4042,
      "mean_token_accuracy": 0.7386852309107781,
      "num_tokens": 378040595.0,
      "step": 5260
    },
    {
      "entropy": 0.970636373758316,
      "epoch": 1.4363586808394657,
      "grad_norm": 0.1145075187087059,
      "learning_rate": 9.942923519569454e-05,
      "loss": 1.4026,
      "mean_token_accuracy": 0.738807100057602,
      "num_tokens": 378756915.0,
      "step": 5270
    },
    {
      "entropy": 0.9653542131185532,
      "epoch": 1.4390842191332789,
      "grad_norm": 0.08687466382980347,
      "learning_rate": 9.94246996423353e-05,
      "loss": 1.407,
      "mean_token_accuracy": 0.7391152828931808,
      "num_tokens": 379472941.0,
      "step": 5280
    },
    {
      "entropy": 0.9694831103086472,
      "epoch": 1.4418097574270918,
      "grad_norm": 0.11827920377254486,
      "learning_rate": 9.942014624372047e-05,
      "loss": 1.403,
      "mean_token_accuracy": 0.7402864143252372,
      "num_tokens": 380199359.0,
      "step": 5290
    },
    {
      "entropy": 0.970561794936657,
      "epoch": 1.4445352957209048,
      "grad_norm": 0.09290741384029388,
      "learning_rate": 9.941557500149413e-05,
      "loss": 1.4052,
      "mean_token_accuracy": 0.7417319104075432,
      "num_tokens": 380929340.0,
      "step": 5300
    },
    {
      "entropy": 0.9641054213047028,
      "epoch": 1.447260834014718,
      "grad_norm": 0.10711231082677841,
      "learning_rate": 9.941098591730674e-05,
      "loss": 1.3992,
      "mean_token_accuracy": 0.7469957336783409,
      "num_tokens": 381668119.0,
      "step": 5310
    },
    {
      "entropy": 0.9665979817509651,
      "epoch": 1.449986372308531,
      "grad_norm": 0.09660854190587997,
      "learning_rate": 9.940637899281524e-05,
      "loss": 1.3924,
      "mean_token_accuracy": 0.7366364017128945,
      "num_tokens": 382384316.0,
      "step": 5320
    },
    {
      "entropy": 0.9691758647561073,
      "epoch": 1.452711910602344,
      "grad_norm": 0.11617083102464676,
      "learning_rate": 9.940175422968301e-05,
      "loss": 1.4093,
      "mean_token_accuracy": 0.7422058954834938,
      "num_tokens": 383106882.0,
      "step": 5330
    },
    {
      "entropy": 0.9698209255933762,
      "epoch": 1.455437448896157,
      "grad_norm": 0.10751962661743164,
      "learning_rate": 9.939711162957987e-05,
      "loss": 1.4099,
      "mean_token_accuracy": 0.7434906512498856,
      "num_tokens": 383833471.0,
      "step": 5340
    },
    {
      "entropy": 0.9739046230912208,
      "epoch": 1.45816298718997,
      "grad_norm": 0.11743414402008057,
      "learning_rate": 9.939245119418207e-05,
      "loss": 1.4089,
      "mean_token_accuracy": 0.7343879967927933,
      "num_tokens": 384534098.0,
      "step": 5350
    },
    {
      "entropy": 0.9686474457383156,
      "epoch": 1.4608885254837831,
      "grad_norm": 0.14783485233783722,
      "learning_rate": 9.93877729251723e-05,
      "loss": 1.401,
      "mean_token_accuracy": 0.7399002060294151,
      "num_tokens": 385258205.0,
      "step": 5360
    },
    {
      "entropy": 0.9598215714097023,
      "epoch": 1.463614063777596,
      "grad_norm": 0.1309163123369217,
      "learning_rate": 9.938307682423971e-05,
      "loss": 1.3964,
      "mean_token_accuracy": 0.7428989082574844,
      "num_tokens": 385970916.0,
      "step": 5370
    },
    {
      "entropy": 0.9654551208019256,
      "epoch": 1.4663396020714092,
      "grad_norm": 0.13030365109443665,
      "learning_rate": 9.937836289307984e-05,
      "loss": 1.3981,
      "mean_token_accuracy": 0.7449974194169044,
      "num_tokens": 386698694.0,
      "step": 5380
    },
    {
      "entropy": 0.9749439135193825,
      "epoch": 1.4690651403652222,
      "grad_norm": 0.12305283546447754,
      "learning_rate": 9.937363113339473e-05,
      "loss": 1.41,
      "mean_token_accuracy": 0.7381033584475517,
      "num_tokens": 387422483.0,
      "step": 5390
    },
    {
      "entropy": 0.9733576580882073,
      "epoch": 1.4717906786590351,
      "grad_norm": 0.09393591433763504,
      "learning_rate": 9.93688815468928e-05,
      "loss": 1.4129,
      "mean_token_accuracy": 0.741550774872303,
      "num_tokens": 388141107.0,
      "step": 5400
    },
    {
      "entropy": 0.9604296833276749,
      "epoch": 1.474516216952848,
      "grad_norm": 0.1078701913356781,
      "learning_rate": 9.936411413528897e-05,
      "loss": 1.3978,
      "mean_token_accuracy": 0.7392246812582016,
      "num_tokens": 388844094.0,
      "step": 5410
    },
    {
      "entropy": 0.9643611937761307,
      "epoch": 1.4772417552466612,
      "grad_norm": 0.15664558112621307,
      "learning_rate": 9.93593289003045e-05,
      "loss": 1.4017,
      "mean_token_accuracy": 0.7436071097850799,
      "num_tokens": 389562699.0,
      "step": 5420
    },
    {
      "entropy": 0.9653159961104393,
      "epoch": 1.4799672935404742,
      "grad_norm": 0.09632451087236404,
      "learning_rate": 9.935452584366722e-05,
      "loss": 1.3946,
      "mean_token_accuracy": 0.7428557753562928,
      "num_tokens": 390288273.0,
      "step": 5430
    },
    {
      "entropy": 0.9551640048623085,
      "epoch": 1.4826928318342874,
      "grad_norm": 0.2352782040834427,
      "learning_rate": 9.934970496711126e-05,
      "loss": 1.3917,
      "mean_token_accuracy": 0.74396021515131,
      "num_tokens": 391025526.0,
      "step": 5440
    },
    {
      "entropy": 0.9665898457169533,
      "epoch": 1.4854183701281003,
      "grad_norm": 0.09772990643978119,
      "learning_rate": 9.934486627237727e-05,
      "loss": 1.3924,
      "mean_token_accuracy": 0.7391120627522468,
      "num_tokens": 391742015.0,
      "step": 5450
    },
    {
      "entropy": 0.971962533891201,
      "epoch": 1.4881439084219132,
      "grad_norm": 0.11168727278709412,
      "learning_rate": 9.93400097612123e-05,
      "loss": 1.408,
      "mean_token_accuracy": 0.73693887591362,
      "num_tokens": 392451387.0,
      "step": 5460
    },
    {
      "entropy": 0.9676396638154984,
      "epoch": 1.4908694467157264,
      "grad_norm": 0.1342191994190216,
      "learning_rate": 9.933513543536983e-05,
      "loss": 1.3999,
      "mean_token_accuracy": 0.7422311305999756,
      "num_tokens": 393168640.0,
      "step": 5470
    },
    {
      "entropy": 0.9660815089941025,
      "epoch": 1.4935949850095394,
      "grad_norm": 0.12180179357528687,
      "learning_rate": 9.933024329660979e-05,
      "loss": 1.3996,
      "mean_token_accuracy": 0.7436567828059196,
      "num_tokens": 393908125.0,
      "step": 5480
    },
    {
      "entropy": 0.972677831351757,
      "epoch": 1.4963205233033525,
      "grad_norm": 0.10382014513015747,
      "learning_rate": 9.932533334669854e-05,
      "loss": 1.4067,
      "mean_token_accuracy": 0.7402523458003998,
      "num_tokens": 394635518.0,
      "step": 5490
    },
    {
      "entropy": 0.9680842772126198,
      "epoch": 1.4990460615971655,
      "grad_norm": 0.10213375091552734,
      "learning_rate": 9.932040558740885e-05,
      "loss": 1.4052,
      "mean_token_accuracy": 0.7380674108862877,
      "num_tokens": 395347447.0,
      "step": 5500
    },
    {
      "entropy": 0.9674480229616165,
      "epoch": 1.5017715998909784,
      "grad_norm": 0.12802939116954803,
      "learning_rate": 9.931546002051995e-05,
      "loss": 1.4073,
      "mean_token_accuracy": 0.7389767229557037,
      "num_tokens": 396055746.0,
      "step": 5510
    },
    {
      "entropy": 0.9710396826267242,
      "epoch": 1.5044971381847914,
      "grad_norm": 0.08243096619844437,
      "learning_rate": 9.931049664781747e-05,
      "loss": 1.4058,
      "mean_token_accuracy": 0.7372735366225243,
      "num_tokens": 396772954.0,
      "step": 5520
    },
    {
      "entropy": 0.9651833549141884,
      "epoch": 1.5072226764786045,
      "grad_norm": 0.09641161561012268,
      "learning_rate": 9.930551547109349e-05,
      "loss": 1.3999,
      "mean_token_accuracy": 0.7431214168667793,
      "num_tokens": 397504248.0,
      "step": 5530
    },
    {
      "entropy": 0.9655771732330323,
      "epoch": 1.5099482147724177,
      "grad_norm": 0.11606141179800034,
      "learning_rate": 9.930051649214651e-05,
      "loss": 1.3927,
      "mean_token_accuracy": 0.7424632519483566,
      "num_tokens": 398237963.0,
      "step": 5540
    },
    {
      "entropy": 0.9724524304270744,
      "epoch": 1.5126737530662306,
      "grad_norm": 0.1653359979391098,
      "learning_rate": 9.929549971278145e-05,
      "loss": 1.413,
      "mean_token_accuracy": 0.7393404498696328,
      "num_tokens": 398958498.0,
      "step": 5550
    },
    {
      "entropy": 0.9636167198419571,
      "epoch": 1.5153992913600436,
      "grad_norm": 0.1128481924533844,
      "learning_rate": 9.929046513480969e-05,
      "loss": 1.3985,
      "mean_token_accuracy": 0.74527777582407,
      "num_tokens": 399699606.0,
      "step": 5560
    },
    {
      "entropy": 0.9632746115326881,
      "epoch": 1.5181248296538565,
      "grad_norm": 0.10375374555587769,
      "learning_rate": 9.928541276004898e-05,
      "loss": 1.4044,
      "mean_token_accuracy": 0.736094568669796,
      "num_tokens": 400392077.0,
      "step": 5570
    },
    {
      "entropy": 0.9567220598459244,
      "epoch": 1.5208503679476697,
      "grad_norm": 0.1052614077925682,
      "learning_rate": 9.928034259032355e-05,
      "loss": 1.4038,
      "mean_token_accuracy": 0.7380430907011032,
      "num_tokens": 401088565.0,
      "step": 5580
    },
    {
      "entropy": 0.9652064919471741,
      "epoch": 1.5235759062414826,
      "grad_norm": 0.10507596284151077,
      "learning_rate": 9.927525462746403e-05,
      "loss": 1.4056,
      "mean_token_accuracy": 0.7403815686702728,
      "num_tokens": 401796270.0,
      "step": 5590
    },
    {
      "entropy": 0.9617900848388672,
      "epoch": 1.5263014445352958,
      "grad_norm": 0.09780077636241913,
      "learning_rate": 9.927014887330746e-05,
      "loss": 1.396,
      "mean_token_accuracy": 0.7425020396709442,
      "num_tokens": 402518111.0,
      "step": 5600
    },
    {
      "entropy": 0.9614929288625718,
      "epoch": 1.5290269828291088,
      "grad_norm": 0.11462049186229706,
      "learning_rate": 9.926502532969733e-05,
      "loss": 1.393,
      "mean_token_accuracy": 0.7415044978260994,
      "num_tokens": 403237289.0,
      "step": 5610
    },
    {
      "entropy": 0.9667704716324806,
      "epoch": 1.5317525211229217,
      "grad_norm": 0.09767358005046844,
      "learning_rate": 9.925988399848356e-05,
      "loss": 1.4032,
      "mean_token_accuracy": 0.7394060298800469,
      "num_tokens": 403952367.0,
      "step": 5620
    },
    {
      "entropy": 0.9641389414668083,
      "epoch": 1.5344780594167347,
      "grad_norm": 0.09655015915632248,
      "learning_rate": 9.925472488152245e-05,
      "loss": 1.399,
      "mean_token_accuracy": 0.7387663036584854,
      "num_tokens": 404658573.0,
      "step": 5630
    },
    {
      "entropy": 0.961449071764946,
      "epoch": 1.5372035977105478,
      "grad_norm": 0.11892081052064896,
      "learning_rate": 9.924954798067678e-05,
      "loss": 1.3976,
      "mean_token_accuracy": 0.7416116803884506,
      "num_tokens": 405381932.0,
      "step": 5640
    },
    {
      "entropy": 0.9588605925440788,
      "epoch": 1.539929136004361,
      "grad_norm": 0.19503852725028992,
      "learning_rate": 9.92443532978157e-05,
      "loss": 1.3907,
      "mean_token_accuracy": 0.7447726503014565,
      "num_tokens": 406116856.0,
      "step": 5650
    },
    {
      "entropy": 0.9709116160869599,
      "epoch": 1.542654674298174,
      "grad_norm": 0.1104922741651535,
      "learning_rate": 9.923914083481479e-05,
      "loss": 1.4042,
      "mean_token_accuracy": 0.7398681297898293,
      "num_tokens": 406830266.0,
      "step": 5660
    },
    {
      "entropy": 0.9636844545602798,
      "epoch": 1.5453802125919869,
      "grad_norm": 0.11828241497278214,
      "learning_rate": 9.923391059355607e-05,
      "loss": 1.3935,
      "mean_token_accuracy": 0.7442167580127717,
      "num_tokens": 407558738.0,
      "step": 5670
    },
    {
      "entropy": 0.9672845751047134,
      "epoch": 1.5481057508857998,
      "grad_norm": 0.10549341887235641,
      "learning_rate": 9.922866257592798e-05,
      "loss": 1.4058,
      "mean_token_accuracy": 0.7364004850387573,
      "num_tokens": 408259142.0,
      "step": 5680
    },
    {
      "entropy": 0.974950310587883,
      "epoch": 1.550831289179613,
      "grad_norm": 0.170731320977211,
      "learning_rate": 9.922339678382533e-05,
      "loss": 1.4041,
      "mean_token_accuracy": 0.7417920991778374,
      "num_tokens": 408985495.0,
      "step": 5690
    },
    {
      "entropy": 0.9687897503376007,
      "epoch": 1.5535568274734262,
      "grad_norm": 0.2253672182559967,
      "learning_rate": 9.921811321914943e-05,
      "loss": 1.3978,
      "mean_token_accuracy": 0.738481967151165,
      "num_tokens": 409705454.0,
      "step": 5700
    },
    {
      "entropy": 0.9635434374213219,
      "epoch": 1.556282365767239,
      "grad_norm": 0.10949309915304184,
      "learning_rate": 9.921281188380793e-05,
      "loss": 1.3972,
      "mean_token_accuracy": 0.7424033030867576,
      "num_tokens": 410425656.0,
      "step": 5710
    },
    {
      "entropy": 0.9592582941055298,
      "epoch": 1.559007904061052,
      "grad_norm": 0.10131136327981949,
      "learning_rate": 9.920749277971493e-05,
      "loss": 1.3983,
      "mean_token_accuracy": 0.7419530689716339,
      "num_tokens": 411136750.0,
      "step": 5720
    },
    {
      "entropy": 0.964599272608757,
      "epoch": 1.561733442354865,
      "grad_norm": 0.10109210014343262,
      "learning_rate": 9.920215590879097e-05,
      "loss": 1.3875,
      "mean_token_accuracy": 0.7396992966532707,
      "num_tokens": 411854916.0,
      "step": 5730
    },
    {
      "entropy": 0.9645843416452408,
      "epoch": 1.5644589806486782,
      "grad_norm": 0.11004949361085892,
      "learning_rate": 9.919680127296294e-05,
      "loss": 1.3946,
      "mean_token_accuracy": 0.7357957765460015,
      "num_tokens": 412563819.0,
      "step": 5740
    },
    {
      "entropy": 0.9718684017658233,
      "epoch": 1.5671845189424911,
      "grad_norm": 0.18444311618804932,
      "learning_rate": 9.919142887416419e-05,
      "loss": 1.4041,
      "mean_token_accuracy": 0.7369261845946312,
      "num_tokens": 413265824.0,
      "step": 5750
    },
    {
      "entropy": 0.9655459359288215,
      "epoch": 1.5699100572363043,
      "grad_norm": 0.10114146769046783,
      "learning_rate": 9.91860387143345e-05,
      "loss": 1.405,
      "mean_token_accuracy": 0.741586534678936,
      "num_tokens": 413987806.0,
      "step": 5760
    },
    {
      "entropy": 0.9581871777772903,
      "epoch": 1.5726355955301172,
      "grad_norm": 0.13442564010620117,
      "learning_rate": 9.918063079542002e-05,
      "loss": 1.393,
      "mean_token_accuracy": 0.7415242746472359,
      "num_tokens": 414707107.0,
      "step": 5770
    },
    {
      "entropy": 0.9689536958932876,
      "epoch": 1.5753611338239302,
      "grad_norm": 0.0971767008304596,
      "learning_rate": 9.917520511937332e-05,
      "loss": 1.4018,
      "mean_token_accuracy": 0.7420008212327958,
      "num_tokens": 415445511.0,
      "step": 5780
    },
    {
      "entropy": 0.9681785985827446,
      "epoch": 1.5780866721177431,
      "grad_norm": 0.11754272133111954,
      "learning_rate": 9.916976168815343e-05,
      "loss": 1.4046,
      "mean_token_accuracy": 0.7410356238484382,
      "num_tokens": 416168666.0,
      "step": 5790
    },
    {
      "entropy": 0.9686637595295906,
      "epoch": 1.5808122104115563,
      "grad_norm": 0.10621481388807297,
      "learning_rate": 9.916430050372572e-05,
      "loss": 1.4034,
      "mean_token_accuracy": 0.7380057781934738,
      "num_tokens": 416880509.0,
      "step": 5800
    },
    {
      "entropy": 0.9566580340266228,
      "epoch": 1.5835377487053695,
      "grad_norm": 0.10499605536460876,
      "learning_rate": 9.9158821568062e-05,
      "loss": 1.3917,
      "mean_token_accuracy": 0.7416859939694405,
      "num_tokens": 417590181.0,
      "step": 5810
    },
    {
      "entropy": 0.9609494879841805,
      "epoch": 1.5862632869991824,
      "grad_norm": 0.09360944479703903,
      "learning_rate": 9.915332488314051e-05,
      "loss": 1.3986,
      "mean_token_accuracy": 0.7389409273862839,
      "num_tokens": 418301686.0,
      "step": 5820
    },
    {
      "entropy": 0.9661891669034958,
      "epoch": 1.5889888252929953,
      "grad_norm": 0.20693224668502808,
      "learning_rate": 9.914781045094586e-05,
      "loss": 1.4011,
      "mean_token_accuracy": 0.7356569826602936,
      "num_tokens": 419002077.0,
      "step": 5830
    },
    {
      "entropy": 0.963114646077156,
      "epoch": 1.5917143635868083,
      "grad_norm": 0.09767451137304306,
      "learning_rate": 9.91422782734691e-05,
      "loss": 1.3977,
      "mean_token_accuracy": 0.7400422096252441,
      "num_tokens": 419715019.0,
      "step": 5840
    },
    {
      "entropy": 0.9726824536919594,
      "epoch": 1.5944399018806215,
      "grad_norm": 0.21406540274620056,
      "learning_rate": 9.913672835270768e-05,
      "loss": 1.4095,
      "mean_token_accuracy": 0.7354395866394043,
      "num_tokens": 420422138.0,
      "step": 5850
    },
    {
      "entropy": 0.9589616864919662,
      "epoch": 1.5971654401744344,
      "grad_norm": 0.20524220168590546,
      "learning_rate": 9.913116069066544e-05,
      "loss": 1.3899,
      "mean_token_accuracy": 0.745594210922718,
      "num_tokens": 421158979.0,
      "step": 5860
    },
    {
      "entropy": 0.9652955681085587,
      "epoch": 1.5998909784682476,
      "grad_norm": 0.10461007803678513,
      "learning_rate": 9.912557528935265e-05,
      "loss": 1.4065,
      "mean_token_accuracy": 0.7428195431828499,
      "num_tokens": 421876794.0,
      "step": 5870
    },
    {
      "entropy": 0.9643538653850555,
      "epoch": 1.6026165167620605,
      "grad_norm": 0.10991375148296356,
      "learning_rate": 9.911997215078595e-05,
      "loss": 1.3987,
      "mean_token_accuracy": 0.7403149724006652,
      "num_tokens": 422592919.0,
      "step": 5880
    },
    {
      "entropy": 0.9613830238580704,
      "epoch": 1.6053420550558735,
      "grad_norm": 0.10263770073652267,
      "learning_rate": 9.911435127698843e-05,
      "loss": 1.3917,
      "mean_token_accuracy": 0.7438402488827706,
      "num_tokens": 423316339.0,
      "step": 5890
    },
    {
      "entropy": 0.9703504428267479,
      "epoch": 1.6080675933496864,
      "grad_norm": 0.10847967118024826,
      "learning_rate": 9.910871266998955e-05,
      "loss": 1.4044,
      "mean_token_accuracy": 0.7379203736782074,
      "num_tokens": 424030583.0,
      "step": 5900
    },
    {
      "entropy": 0.9544369623064994,
      "epoch": 1.6107931316434996,
      "grad_norm": 0.08858538419008255,
      "learning_rate": 9.910305633182518e-05,
      "loss": 1.3887,
      "mean_token_accuracy": 0.7420952394604683,
      "num_tokens": 424760848.0,
      "step": 5910
    },
    {
      "entropy": 0.9609884187579155,
      "epoch": 1.6135186699373127,
      "grad_norm": 0.11026997119188309,
      "learning_rate": 9.90973822645376e-05,
      "loss": 1.3998,
      "mean_token_accuracy": 0.7414646327495575,
      "num_tokens": 425475990.0,
      "step": 5920
    },
    {
      "entropy": 0.9604142621159554,
      "epoch": 1.6162442082311257,
      "grad_norm": 0.10348739475011826,
      "learning_rate": 9.909169047017548e-05,
      "loss": 1.3938,
      "mean_token_accuracy": 0.7371267721056938,
      "num_tokens": 426181074.0,
      "step": 5930
    },
    {
      "entropy": 0.9558591157197952,
      "epoch": 1.6189697465249386,
      "grad_norm": 0.10191468149423599,
      "learning_rate": 9.90859809507939e-05,
      "loss": 1.3941,
      "mean_token_accuracy": 0.7403875023126603,
      "num_tokens": 426902746.0,
      "step": 5940
    },
    {
      "entropy": 0.9646698012948036,
      "epoch": 1.6216952848187516,
      "grad_norm": 0.13848857581615448,
      "learning_rate": 9.908025370845436e-05,
      "loss": 1.4027,
      "mean_token_accuracy": 0.7376966655254364,
      "num_tokens": 427616308.0,
      "step": 5950
    },
    {
      "entropy": 0.9594425529241561,
      "epoch": 1.6244208231125647,
      "grad_norm": 0.1576535850763321,
      "learning_rate": 9.907450874522467e-05,
      "loss": 1.3931,
      "mean_token_accuracy": 0.7456508308649064,
      "num_tokens": 428339961.0,
      "step": 5960
    },
    {
      "entropy": 0.9645897924900055,
      "epoch": 1.627146361406378,
      "grad_norm": 0.1390545815229416,
      "learning_rate": 9.906874606317917e-05,
      "loss": 1.3985,
      "mean_token_accuracy": 0.7383797079324722,
      "num_tokens": 429060230.0,
      "step": 5970
    },
    {
      "entropy": 0.9663798242807389,
      "epoch": 1.6298718997001909,
      "grad_norm": 0.11118265241384506,
      "learning_rate": 9.906296566439848e-05,
      "loss": 1.3904,
      "mean_token_accuracy": 0.7435512930154801,
      "num_tokens": 429800540.0,
      "step": 5980
    },
    {
      "entropy": 0.9650408774614334,
      "epoch": 1.6325974379940038,
      "grad_norm": 0.18296988308429718,
      "learning_rate": 9.905716755096972e-05,
      "loss": 1.3985,
      "mean_token_accuracy": 0.7394289910793305,
      "num_tokens": 430519468.0,
      "step": 5990
    },
    {
      "entropy": 0.9720285251736641,
      "epoch": 1.6353229762878168,
      "grad_norm": 0.10560684651136398,
      "learning_rate": 9.905135172498629e-05,
      "loss": 1.4026,
      "mean_token_accuracy": 0.7365751951932907,
      "num_tokens": 431231969.0,
      "step": 6000
    },
    {
      "entropy": 0.9789033398032189,
      "epoch": 1.63804851458163,
      "grad_norm": 0.12759380042552948,
      "learning_rate": 9.904551818854813e-05,
      "loss": 1.4076,
      "mean_token_accuracy": 0.7391931042075157,
      "num_tokens": 431959887.0,
      "step": 6010
    },
    {
      "entropy": 0.9748961210250855,
      "epoch": 1.6407740528754429,
      "grad_norm": 0.10689742118120193,
      "learning_rate": 9.90396669437614e-05,
      "loss": 1.4072,
      "mean_token_accuracy": 0.7423317298293114,
      "num_tokens": 432679926.0,
      "step": 6020
    },
    {
      "entropy": 0.9654077455401421,
      "epoch": 1.643499591169256,
      "grad_norm": 0.13345809280872345,
      "learning_rate": 9.903379799273884e-05,
      "loss": 1.3972,
      "mean_token_accuracy": 0.7388819694519043,
      "num_tokens": 433407084.0,
      "step": 6030
    },
    {
      "entropy": 0.9621516212821006,
      "epoch": 1.646225129463069,
      "grad_norm": 0.10782342404127121,
      "learning_rate": 9.902791133759943e-05,
      "loss": 1.3989,
      "mean_token_accuracy": 0.741086582839489,
      "num_tokens": 434120698.0,
      "step": 6040
    },
    {
      "entropy": 0.9617944195866585,
      "epoch": 1.648950667756882,
      "grad_norm": 0.1336713284254074,
      "learning_rate": 9.902200698046861e-05,
      "loss": 1.3966,
      "mean_token_accuracy": 0.7426715821027756,
      "num_tokens": 434839891.0,
      "step": 6050
    },
    {
      "entropy": 0.9621335521340371,
      "epoch": 1.6516762060506949,
      "grad_norm": 0.10524009168148041,
      "learning_rate": 9.901608492347823e-05,
      "loss": 1.3959,
      "mean_token_accuracy": 0.740611945092678,
      "num_tokens": 435563923.0,
      "step": 6060
    },
    {
      "entropy": 0.9668157532811165,
      "epoch": 1.654401744344508,
      "grad_norm": 0.11039511114358902,
      "learning_rate": 9.901014516876647e-05,
      "loss": 1.4006,
      "mean_token_accuracy": 0.73515605032444,
      "num_tokens": 436259278.0,
      "step": 6070
    },
    {
      "entropy": 0.9718721225857735,
      "epoch": 1.6571272826383212,
      "grad_norm": 0.09918174147605896,
      "learning_rate": 9.900418771847794e-05,
      "loss": 1.4074,
      "mean_token_accuracy": 0.737083688378334,
      "num_tokens": 436969331.0,
      "step": 6080
    },
    {
      "entropy": 0.9646610051393509,
      "epoch": 1.6598528209321342,
      "grad_norm": 0.21766789257526398,
      "learning_rate": 9.899821257476368e-05,
      "loss": 1.4036,
      "mean_token_accuracy": 0.7371485099196434,
      "num_tokens": 437680886.0,
      "step": 6090
    },
    {
      "entropy": 0.9628597944974899,
      "epoch": 1.662578359225947,
      "grad_norm": 0.11657141149044037,
      "learning_rate": 9.899221973978101e-05,
      "loss": 1.3985,
      "mean_token_accuracy": 0.7447718888521194,
      "num_tokens": 438408284.0,
      "step": 6100
    },
    {
      "entropy": 0.9702211692929268,
      "epoch": 1.66530389751976,
      "grad_norm": 0.10765881091356277,
      "learning_rate": 9.898620921569375e-05,
      "loss": 1.4011,
      "mean_token_accuracy": 0.7367077067494392,
      "num_tokens": 439119517.0,
      "step": 6110
    },
    {
      "entropy": 0.9642821982502937,
      "epoch": 1.6680294358135732,
      "grad_norm": 0.115221306681633,
      "learning_rate": 9.898018100467202e-05,
      "loss": 1.391,
      "mean_token_accuracy": 0.7378931477665901,
      "num_tokens": 439838100.0,
      "step": 6120
    },
    {
      "entropy": 0.9640056222677231,
      "epoch": 1.6707549741073862,
      "grad_norm": 0.10090425610542297,
      "learning_rate": 9.897413510889237e-05,
      "loss": 1.395,
      "mean_token_accuracy": 0.7464444503188133,
      "num_tokens": 440584180.0,
      "step": 6130
    },
    {
      "entropy": 0.9658372700214386,
      "epoch": 1.6734805124011993,
      "grad_norm": 0.14596357941627502,
      "learning_rate": 9.896807153053775e-05,
      "loss": 1.4041,
      "mean_token_accuracy": 0.7407622843980789,
      "num_tokens": 441299995.0,
      "step": 6140
    },
    {
      "entropy": 0.9652875065803528,
      "epoch": 1.6762060506950123,
      "grad_norm": 0.09758126735687256,
      "learning_rate": 9.896199027179744e-05,
      "loss": 1.4041,
      "mean_token_accuracy": 0.7378090232610702,
      "num_tokens": 442012383.0,
      "step": 6150
    },
    {
      "entropy": 0.9647507712244987,
      "epoch": 1.6789315889888252,
      "grad_norm": 0.09823904931545258,
      "learning_rate": 9.895589133486718e-05,
      "loss": 1.3992,
      "mean_token_accuracy": 0.7413588866591454,
      "num_tokens": 442725868.0,
      "step": 6160
    },
    {
      "entropy": 0.9591943696141243,
      "epoch": 1.6816571272826382,
      "grad_norm": 0.08996161073446274,
      "learning_rate": 9.894977472194899e-05,
      "loss": 1.3935,
      "mean_token_accuracy": 0.7463259294629097,
      "num_tokens": 443459707.0,
      "step": 6170
    },
    {
      "entropy": 0.9753120169043541,
      "epoch": 1.6843826655764513,
      "grad_norm": 0.0981547012925148,
      "learning_rate": 9.89436404352514e-05,
      "loss": 1.4098,
      "mean_token_accuracy": 0.7384630024433136,
      "num_tokens": 444177670.0,
      "step": 6180
    },
    {
      "entropy": 0.9743504494428634,
      "epoch": 1.6871082038702645,
      "grad_norm": 0.09443812817335129,
      "learning_rate": 9.893748847698918e-05,
      "loss": 1.4043,
      "mean_token_accuracy": 0.7386202126741409,
      "num_tokens": 444885869.0,
      "step": 6190
    },
    {
      "entropy": 0.9678167894482612,
      "epoch": 1.6898337421640774,
      "grad_norm": 0.09067999571561813,
      "learning_rate": 9.89313188493836e-05,
      "loss": 1.3993,
      "mean_token_accuracy": 0.7457887947559356,
      "num_tokens": 445608265.0,
      "step": 6200
    },
    {
      "entropy": 0.9639017432928085,
      "epoch": 1.6925592804578904,
      "grad_norm": 0.09365516901016235,
      "learning_rate": 9.892513155466226e-05,
      "loss": 1.3901,
      "mean_token_accuracy": 0.7453442245721817,
      "num_tokens": 446340573.0,
      "step": 6210
    },
    {
      "entropy": 0.9653544202446938,
      "epoch": 1.6952848187517033,
      "grad_norm": 0.11369942128658295,
      "learning_rate": 9.891892659505912e-05,
      "loss": 1.398,
      "mean_token_accuracy": 0.7398279547691345,
      "num_tokens": 447053800.0,
      "step": 6220
    },
    {
      "entropy": 0.9619206324219703,
      "epoch": 1.6980103570455165,
      "grad_norm": 0.12091006338596344,
      "learning_rate": 9.891270397281456e-05,
      "loss": 1.3958,
      "mean_token_accuracy": 0.7467226952314376,
      "num_tokens": 447789811.0,
      "step": 6230
    },
    {
      "entropy": 0.9610428914427758,
      "epoch": 1.7007358953393297,
      "grad_norm": 0.09831788390874863,
      "learning_rate": 9.89064636901753e-05,
      "loss": 1.395,
      "mean_token_accuracy": 0.7400000631809235,
      "num_tokens": 448499789.0,
      "step": 6240
    },
    {
      "entropy": 0.9735107973217965,
      "epoch": 1.7034614336331426,
      "grad_norm": 0.14642998576164246,
      "learning_rate": 9.890020574939447e-05,
      "loss": 1.4133,
      "mean_token_accuracy": 0.7345815241336823,
      "num_tokens": 449200840.0,
      "step": 6250
    },
    {
      "entropy": 0.9719688981771469,
      "epoch": 1.7061869719269556,
      "grad_norm": 0.09600794315338135,
      "learning_rate": 9.889393015273153e-05,
      "loss": 1.4037,
      "mean_token_accuracy": 0.741492660343647,
      "num_tokens": 449918387.0,
      "step": 6260
    },
    {
      "entropy": 0.9581882610917092,
      "epoch": 1.7089125102207685,
      "grad_norm": 0.168635755777359,
      "learning_rate": 9.88876369024524e-05,
      "loss": 1.3885,
      "mean_token_accuracy": 0.7411158233880997,
      "num_tokens": 450633441.0,
      "step": 6270
    },
    {
      "entropy": 0.9636551678180695,
      "epoch": 1.7116380485145817,
      "grad_norm": 0.11920435726642609,
      "learning_rate": 9.888132600082925e-05,
      "loss": 1.3973,
      "mean_token_accuracy": 0.7420587435364723,
      "num_tokens": 451351442.0,
      "step": 6280
    },
    {
      "entropy": 0.9653368249535561,
      "epoch": 1.7143635868083946,
      "grad_norm": 0.09253939986228943,
      "learning_rate": 9.887499745014074e-05,
      "loss": 1.4022,
      "mean_token_accuracy": 0.7385347485542297,
      "num_tokens": 452063526.0,
      "step": 6290
    },
    {
      "entropy": 0.9632287070155143,
      "epoch": 1.7170891251022078,
      "grad_norm": 0.09013763070106506,
      "learning_rate": 9.886865125267182e-05,
      "loss": 1.4012,
      "mean_token_accuracy": 0.7379551649093627,
      "num_tokens": 452776328.0,
      "step": 6300
    },
    {
      "entropy": 0.9703898221254349,
      "epoch": 1.7198146633960207,
      "grad_norm": 0.09758932888507843,
      "learning_rate": 9.886228741071386e-05,
      "loss": 1.404,
      "mean_token_accuracy": 0.7333473935723305,
      "num_tokens": 453477490.0,
      "step": 6310
    },
    {
      "entropy": 0.9641054153442383,
      "epoch": 1.7225402016898337,
      "grad_norm": 0.0950767919421196,
      "learning_rate": 9.88559059265646e-05,
      "loss": 1.395,
      "mean_token_accuracy": 0.7399653360247612,
      "num_tokens": 454186713.0,
      "step": 6320
    },
    {
      "entropy": 0.9593479320406914,
      "epoch": 1.7252657399836466,
      "grad_norm": 0.10654161125421524,
      "learning_rate": 9.884950680252811e-05,
      "loss": 1.3914,
      "mean_token_accuracy": 0.7454151093959809,
      "num_tokens": 454921296.0,
      "step": 6330
    },
    {
      "entropy": 0.9649859368801117,
      "epoch": 1.7279912782774598,
      "grad_norm": 0.09915610402822495,
      "learning_rate": 9.884309004091486e-05,
      "loss": 1.393,
      "mean_token_accuracy": 0.742953984439373,
      "num_tokens": 455646589.0,
      "step": 6340
    },
    {
      "entropy": 0.955434288084507,
      "epoch": 1.730716816571273,
      "grad_norm": 0.10013002157211304,
      "learning_rate": 9.88366556440417e-05,
      "loss": 1.3938,
      "mean_token_accuracy": 0.7451090395450592,
      "num_tokens": 456373347.0,
      "step": 6350
    },
    {
      "entropy": 0.9591516688466072,
      "epoch": 1.733442354865086,
      "grad_norm": 0.13232626020908356,
      "learning_rate": 9.883020361423181e-05,
      "loss": 1.3895,
      "mean_token_accuracy": 0.7457116171717644,
      "num_tokens": 457106997.0,
      "step": 6360
    },
    {
      "entropy": 0.9609716460108757,
      "epoch": 1.7361678931588989,
      "grad_norm": 0.1084350049495697,
      "learning_rate": 9.882373395381474e-05,
      "loss": 1.3985,
      "mean_token_accuracy": 0.7440847963094711,
      "num_tokens": 457828350.0,
      "step": 6370
    },
    {
      "entropy": 0.9553082630038261,
      "epoch": 1.7388934314527118,
      "grad_norm": 0.19414696097373962,
      "learning_rate": 9.881724666512649e-05,
      "loss": 1.3913,
      "mean_token_accuracy": 0.7422837510704994,
      "num_tokens": 458539762.0,
      "step": 6380
    },
    {
      "entropy": 0.9681574374437332,
      "epoch": 1.741618969746525,
      "grad_norm": 0.1206468790769577,
      "learning_rate": 9.881074175050928e-05,
      "loss": 1.4007,
      "mean_token_accuracy": 0.7423367232084275,
      "num_tokens": 459265674.0,
      "step": 6390
    },
    {
      "entropy": 0.9651200860738754,
      "epoch": 1.744344508040338,
      "grad_norm": 0.100550577044487,
      "learning_rate": 9.880421921231181e-05,
      "loss": 1.3992,
      "mean_token_accuracy": 0.7399113595485687,
      "num_tokens": 459981436.0,
      "step": 6400
    },
    {
      "entropy": 0.9738900408148765,
      "epoch": 1.747070046334151,
      "grad_norm": 0.11466038227081299,
      "learning_rate": 9.87976790528891e-05,
      "loss": 1.4032,
      "mean_token_accuracy": 0.7395287618041039,
      "num_tokens": 460703872.0,
      "step": 6410
    },
    {
      "entropy": 0.9619228363037109,
      "epoch": 1.749795584627964,
      "grad_norm": 0.253574401140213,
      "learning_rate": 9.879112127460253e-05,
      "loss": 1.3961,
      "mean_token_accuracy": 0.7430710151791573,
      "num_tokens": 461431665.0,
      "step": 6420
    },
    {
      "entropy": 0.9635621175169945,
      "epoch": 1.752521122921777,
      "grad_norm": 0.1853688359260559,
      "learning_rate": 9.878454587981986e-05,
      "loss": 1.3921,
      "mean_token_accuracy": 0.7350759491324425,
      "num_tokens": 462132961.0,
      "step": 6430
    },
    {
      "entropy": 0.9640910357236863,
      "epoch": 1.75524666121559,
      "grad_norm": 0.21477222442626953,
      "learning_rate": 9.877795287091518e-05,
      "loss": 1.4078,
      "mean_token_accuracy": 0.7383334547281265,
      "num_tokens": 462844968.0,
      "step": 6440
    },
    {
      "entropy": 0.9733558312058449,
      "epoch": 1.757972199509403,
      "grad_norm": 0.11411593854427338,
      "learning_rate": 9.877134225026899e-05,
      "loss": 1.4044,
      "mean_token_accuracy": 0.7368480548262596,
      "num_tokens": 463557302.0,
      "step": 6450
    },
    {
      "entropy": 0.9664992153644562,
      "epoch": 1.7606977378032163,
      "grad_norm": 0.11985015124082565,
      "learning_rate": 9.876471402026806e-05,
      "loss": 1.4048,
      "mean_token_accuracy": 0.7415119394659996,
      "num_tokens": 464277878.0,
      "step": 6460
    },
    {
      "entropy": 0.9565216213464737,
      "epoch": 1.7634232760970292,
      "grad_norm": 0.132217139005661,
      "learning_rate": 9.875806818330565e-05,
      "loss": 1.3892,
      "mean_token_accuracy": 0.7427519276738167,
      "num_tokens": 464996856.0,
      "step": 6470
    },
    {
      "entropy": 0.9655373767018318,
      "epoch": 1.7661488143908421,
      "grad_norm": 0.18020625412464142,
      "learning_rate": 9.875140474178123e-05,
      "loss": 1.3953,
      "mean_token_accuracy": 0.7430418714880943,
      "num_tokens": 465723298.0,
      "step": 6480
    },
    {
      "entropy": 0.9673853784799575,
      "epoch": 1.768874352684655,
      "grad_norm": 0.10827246308326721,
      "learning_rate": 9.874472369810074e-05,
      "loss": 1.403,
      "mean_token_accuracy": 0.7388443365693093,
      "num_tokens": 466436406.0,
      "step": 6490
    },
    {
      "entropy": 0.9671308323740959,
      "epoch": 1.7715998909784683,
      "grad_norm": 0.11639074981212616,
      "learning_rate": 9.873802505467645e-05,
      "loss": 1.4035,
      "mean_token_accuracy": 0.7438584089279174,
      "num_tokens": 467165032.0,
      "step": 6500
    },
    {
      "entropy": 0.9688264280557632,
      "epoch": 1.7743254292722814,
      "grad_norm": 0.11113184690475464,
      "learning_rate": 9.873130881392692e-05,
      "loss": 1.4027,
      "mean_token_accuracy": 0.7402560353279114,
      "num_tokens": 467894491.0,
      "step": 6510
    },
    {
      "entropy": 0.9632213458418846,
      "epoch": 1.7770509675660944,
      "grad_norm": 0.10491864383220673,
      "learning_rate": 9.872457497827713e-05,
      "loss": 1.399,
      "mean_token_accuracy": 0.7420696586370468,
      "num_tokens": 468611446.0,
      "step": 6520
    },
    {
      "entropy": 0.9695285797119141,
      "epoch": 1.7797765058599073,
      "grad_norm": 0.10571157187223434,
      "learning_rate": 9.871782355015842e-05,
      "loss": 1.4076,
      "mean_token_accuracy": 0.739043866097927,
      "num_tokens": 469326557.0,
      "step": 6530
    },
    {
      "entropy": 0.9691705301403999,
      "epoch": 1.7825020441537203,
      "grad_norm": 0.27469053864479065,
      "learning_rate": 9.871105453200842e-05,
      "loss": 1.4023,
      "mean_token_accuracy": 0.74033163189888,
      "num_tokens": 470044902.0,
      "step": 6540
    },
    {
      "entropy": 0.9612185657024384,
      "epoch": 1.7852275824475334,
      "grad_norm": 0.19587138295173645,
      "learning_rate": 9.870426792627118e-05,
      "loss": 1.3876,
      "mean_token_accuracy": 0.7447293937206269,
      "num_tokens": 470778600.0,
      "step": 6550
    },
    {
      "entropy": 0.9602941289544106,
      "epoch": 1.7879531207413464,
      "grad_norm": 0.18141105771064758,
      "learning_rate": 9.869746373539707e-05,
      "loss": 1.3949,
      "mean_token_accuracy": 0.7409106940031052,
      "num_tokens": 471495060.0,
      "step": 6560
    },
    {
      "entropy": 0.9630383446812629,
      "epoch": 1.7906786590351595,
      "grad_norm": 0.12980099022388458,
      "learning_rate": 9.869064196184277e-05,
      "loss": 1.3965,
      "mean_token_accuracy": 0.7366779312491417,
      "num_tokens": 472214771.0,
      "step": 6570
    },
    {
      "entropy": 0.9666449546813964,
      "epoch": 1.7934041973289725,
      "grad_norm": 0.14541658759117126,
      "learning_rate": 9.868380260807137e-05,
      "loss": 1.4007,
      "mean_token_accuracy": 0.7406311452388763,
      "num_tokens": 472925946.0,
      "step": 6580
    },
    {
      "entropy": 0.9566129818558693,
      "epoch": 1.7961297356227854,
      "grad_norm": 0.11099447309970856,
      "learning_rate": 9.86769456765523e-05,
      "loss": 1.3948,
      "mean_token_accuracy": 0.7459923014044761,
      "num_tokens": 473657532.0,
      "step": 6590
    },
    {
      "entropy": 0.97569060921669,
      "epoch": 1.7988552739165984,
      "grad_norm": 0.10148832947015762,
      "learning_rate": 9.867007116976128e-05,
      "loss": 1.411,
      "mean_token_accuracy": 0.7368429720401763,
      "num_tokens": 474371122.0,
      "step": 6600
    },
    {
      "entropy": 0.9582658961415291,
      "epoch": 1.8015808122104116,
      "grad_norm": 0.1217406615614891,
      "learning_rate": 9.866317909018046e-05,
      "loss": 1.3888,
      "mean_token_accuracy": 0.7450990483164788,
      "num_tokens": 475096902.0,
      "step": 6610
    },
    {
      "entropy": 0.967861394584179,
      "epoch": 1.8043063505042247,
      "grad_norm": 0.10424516350030899,
      "learning_rate": 9.865626944029825e-05,
      "loss": 1.3986,
      "mean_token_accuracy": 0.740696033835411,
      "num_tokens": 475810807.0,
      "step": 6620
    },
    {
      "entropy": 0.9663902729749679,
      "epoch": 1.8070318887980377,
      "grad_norm": 0.0985787957906723,
      "learning_rate": 9.864934222260946e-05,
      "loss": 1.4122,
      "mean_token_accuracy": 0.7326736062765121,
      "num_tokens": 476500759.0,
      "step": 6630
    },
    {
      "entropy": 0.9576504468917847,
      "epoch": 1.8097574270918506,
      "grad_norm": 0.10975080728530884,
      "learning_rate": 9.864239743961526e-05,
      "loss": 1.3921,
      "mean_token_accuracy": 0.7430121034383774,
      "num_tokens": 477217083.0,
      "step": 6640
    },
    {
      "entropy": 0.9580236986279488,
      "epoch": 1.8124829653856636,
      "grad_norm": 0.12971186637878418,
      "learning_rate": 9.863543509382306e-05,
      "loss": 1.388,
      "mean_token_accuracy": 0.7417103409767151,
      "num_tokens": 477929836.0,
      "step": 6650
    },
    {
      "entropy": 0.9550999134778977,
      "epoch": 1.8152085036794767,
      "grad_norm": 0.11257505416870117,
      "learning_rate": 9.862845518774675e-05,
      "loss": 1.3906,
      "mean_token_accuracy": 0.7463659942150116,
      "num_tokens": 478663354.0,
      "step": 6660
    },
    {
      "entropy": 0.9664853855967521,
      "epoch": 1.8179340419732897,
      "grad_norm": 0.15594874322414398,
      "learning_rate": 9.862145772390646e-05,
      "loss": 1.4018,
      "mean_token_accuracy": 0.7420924484729767,
      "num_tokens": 479385914.0,
      "step": 6670
    },
    {
      "entropy": 0.9597770988941192,
      "epoch": 1.8206595802671028,
      "grad_norm": 0.0977451428771019,
      "learning_rate": 9.861444270482868e-05,
      "loss": 1.3907,
      "mean_token_accuracy": 0.7440643236041069,
      "num_tokens": 480110862.0,
      "step": 6680
    },
    {
      "entropy": 0.9581946909427643,
      "epoch": 1.8233851185609158,
      "grad_norm": 0.19408464431762695,
      "learning_rate": 9.860741013304626e-05,
      "loss": 1.4026,
      "mean_token_accuracy": 0.7428516134619713,
      "num_tokens": 480828950.0,
      "step": 6690
    },
    {
      "entropy": 0.9585664570331573,
      "epoch": 1.8261106568547287,
      "grad_norm": 0.10100160539150238,
      "learning_rate": 9.860036001109839e-05,
      "loss": 1.3887,
      "mean_token_accuracy": 0.7387926205992699,
      "num_tokens": 481542639.0,
      "step": 6700
    },
    {
      "entropy": 0.9834058821201325,
      "epoch": 1.8288361951485417,
      "grad_norm": 0.09401005506515503,
      "learning_rate": 9.859329234153056e-05,
      "loss": 1.4225,
      "mean_token_accuracy": 0.7341606616973877,
      "num_tokens": 482246841.0,
      "step": 6710
    },
    {
      "entropy": 0.9572040036320686,
      "epoch": 1.8315617334423548,
      "grad_norm": 0.17733122408390045,
      "learning_rate": 9.858620712689464e-05,
      "loss": 1.386,
      "mean_token_accuracy": 0.7434762373566628,
      "num_tokens": 482975640.0,
      "step": 6720
    },
    {
      "entropy": 0.9656352862715721,
      "epoch": 1.834287271736168,
      "grad_norm": 0.1350366473197937,
      "learning_rate": 9.85791043697488e-05,
      "loss": 1.4004,
      "mean_token_accuracy": 0.7439049080014228,
      "num_tokens": 483706489.0,
      "step": 6730
    },
    {
      "entropy": 0.9673403173685073,
      "epoch": 1.837012810029981,
      "grad_norm": 0.10896478593349457,
      "learning_rate": 9.857198407265757e-05,
      "loss": 1.3988,
      "mean_token_accuracy": 0.7389033928513526,
      "num_tokens": 484413613.0,
      "step": 6740
    },
    {
      "entropy": 0.9576301813125611,
      "epoch": 1.839738348323794,
      "grad_norm": 0.09817972779273987,
      "learning_rate": 9.856484623819177e-05,
      "loss": 1.3967,
      "mean_token_accuracy": 0.7377318128943443,
      "num_tokens": 485121971.0,
      "step": 6750
    },
    {
      "entropy": 0.9576287031173706,
      "epoch": 1.8424638866176068,
      "grad_norm": 0.08615750819444656,
      "learning_rate": 9.855769086892863e-05,
      "loss": 1.3934,
      "mean_token_accuracy": 0.7438340544700622,
      "num_tokens": 485844189.0,
      "step": 6760
    },
    {
      "entropy": 0.965782268345356,
      "epoch": 1.84518942491142,
      "grad_norm": 0.0871151015162468,
      "learning_rate": 9.855051796745163e-05,
      "loss": 1.4058,
      "mean_token_accuracy": 0.7387169077992439,
      "num_tokens": 486551935.0,
      "step": 6770
    },
    {
      "entropy": 0.9585666686296463,
      "epoch": 1.8479149632052332,
      "grad_norm": 0.09674597531557083,
      "learning_rate": 9.854332753635063e-05,
      "loss": 1.3932,
      "mean_token_accuracy": 0.7405648306012154,
      "num_tokens": 487268612.0,
      "step": 6780
    },
    {
      "entropy": 0.965125958621502,
      "epoch": 1.8506405014990461,
      "grad_norm": 0.12495724111795425,
      "learning_rate": 9.853611957822177e-05,
      "loss": 1.3854,
      "mean_token_accuracy": 0.7433117389678955,
      "num_tokens": 487996867.0,
      "step": 6790
    },
    {
      "entropy": 0.9590855464339256,
      "epoch": 1.853366039792859,
      "grad_norm": 0.11312315613031387,
      "learning_rate": 9.852889409566761e-05,
      "loss": 1.3957,
      "mean_token_accuracy": 0.746566678583622,
      "num_tokens": 488725261.0,
      "step": 6800
    },
    {
      "entropy": 0.9670864477753639,
      "epoch": 1.856091578086672,
      "grad_norm": 0.2885074019432068,
      "learning_rate": 9.852165109129693e-05,
      "loss": 1.4026,
      "mean_token_accuracy": 0.742882926762104,
      "num_tokens": 489451203.0,
      "step": 6810
    },
    {
      "entropy": 0.9638143733143807,
      "epoch": 1.8588171163804852,
      "grad_norm": 0.1008826196193695,
      "learning_rate": 9.85143905677249e-05,
      "loss": 1.3971,
      "mean_token_accuracy": 0.742574380338192,
      "num_tokens": 490162571.0,
      "step": 6820
    },
    {
      "entropy": 0.9604418158531189,
      "epoch": 1.8615426546742981,
      "grad_norm": 0.11003129184246063,
      "learning_rate": 9.8507112527573e-05,
      "loss": 1.4072,
      "mean_token_accuracy": 0.7377557650208473,
      "num_tokens": 490867938.0,
      "step": 6830
    },
    {
      "entropy": 0.9772702425718307,
      "epoch": 1.8642681929681113,
      "grad_norm": 0.11208124458789825,
      "learning_rate": 9.849981697346905e-05,
      "loss": 1.4098,
      "mean_token_accuracy": 0.7375947162508965,
      "num_tokens": 491576129.0,
      "step": 6840
    },
    {
      "entropy": 0.9600740447640419,
      "epoch": 1.8669937312619242,
      "grad_norm": 0.09464003145694733,
      "learning_rate": 9.849250390804716e-05,
      "loss": 1.3939,
      "mean_token_accuracy": 0.7422061085700988,
      "num_tokens": 492307063.0,
      "step": 6850
    },
    {
      "entropy": 0.968334223330021,
      "epoch": 1.8697192695557372,
      "grad_norm": 0.0899537205696106,
      "learning_rate": 9.848517333394778e-05,
      "loss": 1.3974,
      "mean_token_accuracy": 0.7402063041925431,
      "num_tokens": 493034976.0,
      "step": 6860
    },
    {
      "entropy": 0.9650499865412712,
      "epoch": 1.8724448078495501,
      "grad_norm": 0.12556807696819305,
      "learning_rate": 9.847782525381771e-05,
      "loss": 1.3854,
      "mean_token_accuracy": 0.7403883025050163,
      "num_tokens": 493751088.0,
      "step": 6870
    },
    {
      "entropy": 0.9697022259235382,
      "epoch": 1.8751703461433633,
      "grad_norm": 0.0861344039440155,
      "learning_rate": 9.847045967031e-05,
      "loss": 1.4043,
      "mean_token_accuracy": 0.7371412217617035,
      "num_tokens": 494470261.0,
      "step": 6880
    },
    {
      "entropy": 0.9598774820566177,
      "epoch": 1.8778958844371765,
      "grad_norm": 0.2520541548728943,
      "learning_rate": 9.846307658608411e-05,
      "loss": 1.3922,
      "mean_token_accuracy": 0.7408092856407166,
      "num_tokens": 495185157.0,
      "step": 6890
    },
    {
      "entropy": 0.9653854072093964,
      "epoch": 1.8806214227309894,
      "grad_norm": 0.09129787236452103,
      "learning_rate": 9.845567600380577e-05,
      "loss": 1.4025,
      "mean_token_accuracy": 0.7407778039574623,
      "num_tokens": 495901658.0,
      "step": 6900
    },
    {
      "entropy": 0.9657075956463814,
      "epoch": 1.8833469610248024,
      "grad_norm": 0.11199197173118591,
      "learning_rate": 9.844825792614698e-05,
      "loss": 1.3948,
      "mean_token_accuracy": 0.7429806426167488,
      "num_tokens": 496613845.0,
      "step": 6910
    },
    {
      "entropy": 0.9662411883473396,
      "epoch": 1.8860724993186153,
      "grad_norm": 0.10727035999298096,
      "learning_rate": 9.844082235578613e-05,
      "loss": 1.4028,
      "mean_token_accuracy": 0.7396739244461059,
      "num_tokens": 497323165.0,
      "step": 6920
    },
    {
      "entropy": 0.9622104436159133,
      "epoch": 1.8887980376124285,
      "grad_norm": 0.09176908433437347,
      "learning_rate": 9.843336929540794e-05,
      "loss": 1.3901,
      "mean_token_accuracy": 0.7427182823419571,
      "num_tokens": 498051189.0,
      "step": 6930
    },
    {
      "entropy": 0.9555501505732537,
      "epoch": 1.8915235759062414,
      "grad_norm": 0.0917062908411026,
      "learning_rate": 9.842589874770336e-05,
      "loss": 1.3965,
      "mean_token_accuracy": 0.7373476296663284,
      "num_tokens": 498746121.0,
      "step": 6940
    },
    {
      "entropy": 0.9639024570584297,
      "epoch": 1.8942491142000546,
      "grad_norm": 0.11429794877767563,
      "learning_rate": 9.841841071536973e-05,
      "loss": 1.3835,
      "mean_token_accuracy": 0.7447345286607743,
      "num_tokens": 499477793.0,
      "step": 6950
    },
    {
      "entropy": 0.9644086465239525,
      "epoch": 1.8969746524938675,
      "grad_norm": 0.11062592267990112,
      "learning_rate": 9.841090520111067e-05,
      "loss": 1.3984,
      "mean_token_accuracy": 0.7412375822663307,
      "num_tokens": 500197918.0,
      "step": 6960
    },
    {
      "entropy": 0.963076826930046,
      "epoch": 1.8997001907876805,
      "grad_norm": 0.10753937810659409,
      "learning_rate": 9.84033822076361e-05,
      "loss": 1.3964,
      "mean_token_accuracy": 0.7429899856448173,
      "num_tokens": 500926447.0,
      "step": 6970
    },
    {
      "entropy": 0.9709880262613296,
      "epoch": 1.9024257290814934,
      "grad_norm": 0.10585269331932068,
      "learning_rate": 9.83958417376623e-05,
      "loss": 1.4053,
      "mean_token_accuracy": 0.7401735603809356,
      "num_tokens": 501640460.0,
      "step": 6980
    },
    {
      "entropy": 0.9630426838994026,
      "epoch": 1.9051512673753066,
      "grad_norm": 0.10134958475828171,
      "learning_rate": 9.838828379391181e-05,
      "loss": 1.3888,
      "mean_token_accuracy": 0.7396430373191833,
      "num_tokens": 502365090.0,
      "step": 6990
    },
    {
      "entropy": 0.9591939330101014,
      "epoch": 1.9078768056691198,
      "grad_norm": 0.09174404293298721,
      "learning_rate": 9.838070837911348e-05,
      "loss": 1.3952,
      "mean_token_accuracy": 0.7454899072647094,
      "num_tokens": 503084759.0,
      "step": 7000
    },
    {
      "entropy": 0.9618557408452034,
      "epoch": 1.9106023439629327,
      "grad_norm": 0.09564127027988434,
      "learning_rate": 9.837311549600249e-05,
      "loss": 1.4034,
      "mean_token_accuracy": 0.7399124518036843,
      "num_tokens": 503798396.0,
      "step": 7010
    },
    {
      "entropy": 0.96431335657835,
      "epoch": 1.9133278822567457,
      "grad_norm": 0.10930278897285461,
      "learning_rate": 9.836550514732037e-05,
      "loss": 1.4017,
      "mean_token_accuracy": 0.7392303958535195,
      "num_tokens": 504497386.0,
      "step": 7020
    },
    {
      "entropy": 0.9633992880582809,
      "epoch": 1.9160534205505586,
      "grad_norm": 0.1634802371263504,
      "learning_rate": 9.835787733581486e-05,
      "loss": 1.3977,
      "mean_token_accuracy": 0.7398845091462135,
      "num_tokens": 505215846.0,
      "step": 7030
    },
    {
      "entropy": 0.9664142698049545,
      "epoch": 1.9187789588443718,
      "grad_norm": 0.09212366491556168,
      "learning_rate": 9.835023206424005e-05,
      "loss": 1.3979,
      "mean_token_accuracy": 0.7405804008245468,
      "num_tokens": 505943557.0,
      "step": 7040
    },
    {
      "entropy": 0.9642082065343857,
      "epoch": 1.921504497138185,
      "grad_norm": 0.09991949796676636,
      "learning_rate": 9.834256933535637e-05,
      "loss": 1.4027,
      "mean_token_accuracy": 0.7385505959391594,
      "num_tokens": 506652597.0,
      "step": 7050
    },
    {
      "entropy": 0.9579337358474731,
      "epoch": 1.9242300354319979,
      "grad_norm": 0.08951427787542343,
      "learning_rate": 9.833488915193049e-05,
      "loss": 1.3837,
      "mean_token_accuracy": 0.745193175971508,
      "num_tokens": 507385479.0,
      "step": 7060
    },
    {
      "entropy": 0.9568658128380776,
      "epoch": 1.9269555737258108,
      "grad_norm": 0.18916438519954681,
      "learning_rate": 9.832719151673545e-05,
      "loss": 1.3848,
      "mean_token_accuracy": 0.7461760938167572,
      "num_tokens": 508128269.0,
      "step": 7070
    },
    {
      "entropy": 0.95683603733778,
      "epoch": 1.9296811120196238,
      "grad_norm": 0.12292623519897461,
      "learning_rate": 9.831947643255053e-05,
      "loss": 1.3861,
      "mean_token_accuracy": 0.7436719551682472,
      "num_tokens": 508847491.0,
      "step": 7080
    },
    {
      "entropy": 0.9681815043091774,
      "epoch": 1.932406650313437,
      "grad_norm": 0.12597902119159698,
      "learning_rate": 9.831174390216131e-05,
      "loss": 1.3973,
      "mean_token_accuracy": 0.7405972227454185,
      "num_tokens": 509565726.0,
      "step": 7090
    },
    {
      "entropy": 0.9576827391982079,
      "epoch": 1.93513218860725,
      "grad_norm": 0.13779397308826447,
      "learning_rate": 9.830399392835974e-05,
      "loss": 1.3894,
      "mean_token_accuracy": 0.744329534471035,
      "num_tokens": 510287077.0,
      "step": 7100
    },
    {
      "entropy": 0.9657632574439049,
      "epoch": 1.937857726901063,
      "grad_norm": 0.1731322556734085,
      "learning_rate": 9.829622651394399e-05,
      "loss": 1.4007,
      "mean_token_accuracy": 0.739681999385357,
      "num_tokens": 511002391.0,
      "step": 7110
    },
    {
      "entropy": 0.9532843604683876,
      "epoch": 1.940583265194876,
      "grad_norm": 0.11684025079011917,
      "learning_rate": 9.828844166171858e-05,
      "loss": 1.3901,
      "mean_token_accuracy": 0.745066399872303,
      "num_tokens": 511722117.0,
      "step": 7120
    },
    {
      "entropy": 0.9625393569469451,
      "epoch": 1.943308803488689,
      "grad_norm": 0.09408530592918396,
      "learning_rate": 9.828063937449428e-05,
      "loss": 1.3974,
      "mean_token_accuracy": 0.7381411388516426,
      "num_tokens": 512430728.0,
      "step": 7130
    },
    {
      "entropy": 0.9637985542416573,
      "epoch": 1.946034341782502,
      "grad_norm": 0.09656120091676712,
      "learning_rate": 9.827281965508821e-05,
      "loss": 1.3921,
      "mean_token_accuracy": 0.7378409832715989,
      "num_tokens": 513136482.0,
      "step": 7140
    },
    {
      "entropy": 0.9575431600213051,
      "epoch": 1.948759880076315,
      "grad_norm": 0.122360460460186,
      "learning_rate": 9.826498250632373e-05,
      "loss": 1.3869,
      "mean_token_accuracy": 0.7458069577813149,
      "num_tokens": 513856276.0,
      "step": 7150
    },
    {
      "entropy": 0.9583240255713463,
      "epoch": 1.9514854183701282,
      "grad_norm": 0.12442438304424286,
      "learning_rate": 9.82571279310305e-05,
      "loss": 1.3951,
      "mean_token_accuracy": 0.7432518944144249,
      "num_tokens": 514579908.0,
      "step": 7160
    },
    {
      "entropy": 0.9575931891798973,
      "epoch": 1.9542109566639412,
      "grad_norm": 0.11884919553995132,
      "learning_rate": 9.824925593204455e-05,
      "loss": 1.3841,
      "mean_token_accuracy": 0.7429573744535446,
      "num_tokens": 515296332.0,
      "step": 7170
    },
    {
      "entropy": 0.9661365717649459,
      "epoch": 1.9569364949577541,
      "grad_norm": 0.11476761847734451,
      "learning_rate": 9.824136651220807e-05,
      "loss": 1.3997,
      "mean_token_accuracy": 0.7430660113692283,
      "num_tokens": 516010205.0,
      "step": 7180
    },
    {
      "entropy": 0.956727659702301,
      "epoch": 1.959662033251567,
      "grad_norm": 0.10957435518503189,
      "learning_rate": 9.823345967436967e-05,
      "loss": 1.3944,
      "mean_token_accuracy": 0.7411398693919182,
      "num_tokens": 516731302.0,
      "step": 7190
    },
    {
      "entropy": 0.9655549123883247,
      "epoch": 1.9623875715453802,
      "grad_norm": 0.09681804478168488,
      "learning_rate": 9.822553542138415e-05,
      "loss": 1.3992,
      "mean_token_accuracy": 0.7471737802028656,
      "num_tokens": 517470113.0,
      "step": 7200
    },
    {
      "entropy": 0.9667171403765679,
      "epoch": 1.9651131098391934,
      "grad_norm": 0.0997747853398323,
      "learning_rate": 9.821759375611263e-05,
      "loss": 1.403,
      "mean_token_accuracy": 0.737883883714676,
      "num_tokens": 518174523.0,
      "step": 7210
    },
    {
      "entropy": 0.9553831577301025,
      "epoch": 1.9678386481330064,
      "grad_norm": 0.11053860187530518,
      "learning_rate": 9.820963468142257e-05,
      "loss": 1.3945,
      "mean_token_accuracy": 0.7452098578214645,
      "num_tokens": 518902470.0,
      "step": 7220
    },
    {
      "entropy": 0.9650885179638863,
      "epoch": 1.9705641864268193,
      "grad_norm": 0.2715238630771637,
      "learning_rate": 9.820165820018762e-05,
      "loss": 1.3949,
      "mean_token_accuracy": 0.7417281478643417,
      "num_tokens": 519633955.0,
      "step": 7230
    },
    {
      "entropy": 0.9656887158751488,
      "epoch": 1.9732897247206322,
      "grad_norm": 0.1260688304901123,
      "learning_rate": 9.819366431528778e-05,
      "loss": 1.3988,
      "mean_token_accuracy": 0.7414224296808243,
      "num_tokens": 520353696.0,
      "step": 7240
    },
    {
      "entropy": 0.9694205418229103,
      "epoch": 1.9760152630144452,
      "grad_norm": 0.10043631494045258,
      "learning_rate": 9.818565302960933e-05,
      "loss": 1.4004,
      "mean_token_accuracy": 0.738286642730236,
      "num_tokens": 521068351.0,
      "step": 7250
    },
    {
      "entropy": 0.9686072587966919,
      "epoch": 1.9787408013082584,
      "grad_norm": 0.10541510581970215,
      "learning_rate": 9.817762434604481e-05,
      "loss": 1.3909,
      "mean_token_accuracy": 0.7393988057971,
      "num_tokens": 521795139.0,
      "step": 7260
    },
    {
      "entropy": 0.9675464764237404,
      "epoch": 1.9814663396020715,
      "grad_norm": 0.11267922073602676,
      "learning_rate": 9.816957826749307e-05,
      "loss": 1.3967,
      "mean_token_accuracy": 0.7457544431090355,
      "num_tokens": 522520008.0,
      "step": 7270
    },
    {
      "entropy": 0.9636656671762467,
      "epoch": 1.9841918778958845,
      "grad_norm": 0.21923284232616425,
      "learning_rate": 9.816151479685918e-05,
      "loss": 1.408,
      "mean_token_accuracy": 0.7386617347598076,
      "num_tokens": 523232215.0,
      "step": 7280
    },
    {
      "entropy": 0.9583509057760239,
      "epoch": 1.9869174161896974,
      "grad_norm": 0.10762002319097519,
      "learning_rate": 9.815343393705458e-05,
      "loss": 1.3984,
      "mean_token_accuracy": 0.7422628074884414,
      "num_tokens": 523954702.0,
      "step": 7290
    },
    {
      "entropy": 0.9607061296701431,
      "epoch": 1.9896429544835104,
      "grad_norm": 0.150059774518013,
      "learning_rate": 9.814533569099691e-05,
      "loss": 1.4031,
      "mean_token_accuracy": 0.7383078336715698,
      "num_tokens": 524655836.0,
      "step": 7300
    },
    {
      "entropy": 0.9600534334778785,
      "epoch": 1.9923684927773235,
      "grad_norm": 0.14201506972312927,
      "learning_rate": 9.813722006161013e-05,
      "loss": 1.3871,
      "mean_token_accuracy": 0.743747016787529,
      "num_tokens": 525384534.0,
      "step": 7310
    },
    {
      "entropy": 0.9653302758932114,
      "epoch": 1.9950940310711367,
      "grad_norm": 0.14192071557044983,
      "learning_rate": 9.812908705182447e-05,
      "loss": 1.4038,
      "mean_token_accuracy": 0.7368806630373002,
      "num_tokens": 526086698.0,
      "step": 7320
    },
    {
      "entropy": 0.9466646060347557,
      "epoch": 1.9978195693649496,
      "grad_norm": 0.10055558383464813,
      "learning_rate": 9.81209366645764e-05,
      "loss": 1.3775,
      "mean_token_accuracy": 0.7499463304877281,
      "num_tokens": 526816365.0,
      "step": 7330
    },
    {
      "entropy": 0.9586656853556633,
      "epoch": 2.0005451076587626,
      "grad_norm": 0.10854324698448181,
      "learning_rate": 9.811276890280874e-05,
      "loss": 1.3988,
      "mean_token_accuracy": 0.7433884471654892,
      "num_tokens": 527542733.0,
      "step": 7340
    },
    {
      "entropy": 0.9561543703079224,
      "epoch": 2.0032706459525755,
      "grad_norm": 0.1269897073507309,
      "learning_rate": 9.810458376947052e-05,
      "loss": 1.387,
      "mean_token_accuracy": 0.7402018457651138,
      "num_tokens": 528249256.0,
      "step": 7350
    },
    {
      "entropy": 0.9542464792728425,
      "epoch": 2.0059961842463885,
      "grad_norm": 0.15228229761123657,
      "learning_rate": 9.809638126751703e-05,
      "loss": 1.3788,
      "mean_token_accuracy": 0.7452547445893287,
      "num_tokens": 528984076.0,
      "step": 7360
    },
    {
      "entropy": 0.960684311389923,
      "epoch": 2.008721722540202,
      "grad_norm": 0.17953327298164368,
      "learning_rate": 9.80881613999099e-05,
      "loss": 1.3918,
      "mean_token_accuracy": 0.7386097416281701,
      "num_tokens": 529690825.0,
      "step": 7370
    },
    {
      "entropy": 0.963483664393425,
      "epoch": 2.011447260834015,
      "grad_norm": 0.13395299017429352,
      "learning_rate": 9.8079924169617e-05,
      "loss": 1.3908,
      "mean_token_accuracy": 0.7427313566207886,
      "num_tokens": 530403055.0,
      "step": 7380
    },
    {
      "entropy": 0.9560441017150879,
      "epoch": 2.0141727991278278,
      "grad_norm": 0.09408120065927505,
      "learning_rate": 9.807166957961241e-05,
      "loss": 1.3893,
      "mean_token_accuracy": 0.745306758582592,
      "num_tokens": 531136596.0,
      "step": 7390
    },
    {
      "entropy": 0.951741099357605,
      "epoch": 2.0168983374216407,
      "grad_norm": 0.14150190353393555,
      "learning_rate": 9.806339763287656e-05,
      "loss": 1.3837,
      "mean_token_accuracy": 0.7447185799479484,
      "num_tokens": 531852880.0,
      "step": 7400
    },
    {
      "entropy": 0.9562306642532349,
      "epoch": 2.0196238757154537,
      "grad_norm": 0.10243608802556992,
      "learning_rate": 9.805510833239612e-05,
      "loss": 1.3934,
      "mean_token_accuracy": 0.741636997461319,
      "num_tokens": 532563052.0,
      "step": 7410
    },
    {
      "entropy": 0.9499559327960014,
      "epoch": 2.022349414009267,
      "grad_norm": 0.10605795681476593,
      "learning_rate": 9.8046801681164e-05,
      "loss": 1.3769,
      "mean_token_accuracy": 0.7456878125667572,
      "num_tokens": 533291538.0,
      "step": 7420
    },
    {
      "entropy": 0.9550616160035134,
      "epoch": 2.02507495230308,
      "grad_norm": 0.10388023406267166,
      "learning_rate": 9.803847768217941e-05,
      "loss": 1.383,
      "mean_token_accuracy": 0.7446332737803459,
      "num_tokens": 534021094.0,
      "step": 7430
    },
    {
      "entropy": 0.9521099284291268,
      "epoch": 2.027800490596893,
      "grad_norm": 0.08860861510038376,
      "learning_rate": 9.80301363384478e-05,
      "loss": 1.3937,
      "mean_token_accuracy": 0.7413805097341537,
      "num_tokens": 534732981.0,
      "step": 7440
    },
    {
      "entropy": 0.9576058641076088,
      "epoch": 2.030526028890706,
      "grad_norm": 0.10883308202028275,
      "learning_rate": 9.802177765298091e-05,
      "loss": 1.3906,
      "mean_token_accuracy": 0.7408022314310074,
      "num_tokens": 535446088.0,
      "step": 7450
    },
    {
      "entropy": 0.9478833019733429,
      "epoch": 2.033251567184519,
      "grad_norm": 0.12634532153606415,
      "learning_rate": 9.801340162879669e-05,
      "loss": 1.3839,
      "mean_token_accuracy": 0.7444364428520203,
      "num_tokens": 536164628.0,
      "step": 7460
    },
    {
      "entropy": 0.9460383102297782,
      "epoch": 2.0359771054783318,
      "grad_norm": 0.09973424673080444,
      "learning_rate": 9.800500826891941e-05,
      "loss": 1.38,
      "mean_token_accuracy": 0.7519521161913871,
      "num_tokens": 536907046.0,
      "step": 7470
    },
    {
      "entropy": 0.966365572810173,
      "epoch": 2.038702643772145,
      "grad_norm": 0.08847790956497192,
      "learning_rate": 9.799659757637956e-05,
      "loss": 1.3919,
      "mean_token_accuracy": 0.7421518012881279,
      "num_tokens": 537631637.0,
      "step": 7480
    },
    {
      "entropy": 0.9592581987380981,
      "epoch": 2.041428182065958,
      "grad_norm": 0.09009517729282379,
      "learning_rate": 9.798816955421389e-05,
      "loss": 1.3882,
      "mean_token_accuracy": 0.7410011321306229,
      "num_tokens": 538358781.0,
      "step": 7490
    },
    {
      "entropy": 0.9531476750969887,
      "epoch": 2.044153720359771,
      "grad_norm": 0.09477727115154266,
      "learning_rate": 9.797972420546544e-05,
      "loss": 1.3776,
      "mean_token_accuracy": 0.7424200236797333,
      "num_tokens": 539067922.0,
      "step": 7500
    },
    {
      "entropy": 0.9474761784076691,
      "epoch": 2.046879258653584,
      "grad_norm": 0.09691348671913147,
      "learning_rate": 9.797126153318348e-05,
      "loss": 1.3829,
      "mean_token_accuracy": 0.7442451179027557,
      "num_tokens": 539778951.0,
      "step": 7510
    },
    {
      "entropy": 0.9564800217747689,
      "epoch": 2.049604796947397,
      "grad_norm": 0.09522726386785507,
      "learning_rate": 9.796278154042353e-05,
      "loss": 1.3926,
      "mean_token_accuracy": 0.7425307676196098,
      "num_tokens": 540497847.0,
      "step": 7520
    },
    {
      "entropy": 0.9547890648245811,
      "epoch": 2.0523303352412103,
      "grad_norm": 0.14827516674995422,
      "learning_rate": 9.795428423024736e-05,
      "loss": 1.3885,
      "mean_token_accuracy": 0.7409161731600762,
      "num_tokens": 541210115.0,
      "step": 7530
    },
    {
      "entropy": 0.9494491994380951,
      "epoch": 2.0550558735350233,
      "grad_norm": 0.10479316115379333,
      "learning_rate": 9.794576960572302e-05,
      "loss": 1.3885,
      "mean_token_accuracy": 0.739873006939888,
      "num_tokens": 541913200.0,
      "step": 7540
    },
    {
      "entropy": 0.9518936529755593,
      "epoch": 2.0577814118288362,
      "grad_norm": 0.1043408066034317,
      "learning_rate": 9.793723766992479e-05,
      "loss": 1.3808,
      "mean_token_accuracy": 0.7446469202637672,
      "num_tokens": 542633892.0,
      "step": 7550
    },
    {
      "entropy": 0.9660930693149566,
      "epoch": 2.060506950122649,
      "grad_norm": 0.11151346564292908,
      "learning_rate": 9.792868842593322e-05,
      "loss": 1.3953,
      "mean_token_accuracy": 0.7388115391135216,
      "num_tokens": 543356038.0,
      "step": 7560
    },
    {
      "entropy": 0.9554453119635582,
      "epoch": 2.063232488416462,
      "grad_norm": 0.10359987616539001,
      "learning_rate": 9.792012187683508e-05,
      "loss": 1.3899,
      "mean_token_accuracy": 0.7420963943004608,
      "num_tokens": 544074279.0,
      "step": 7570
    },
    {
      "entropy": 0.9530102029442787,
      "epoch": 2.065958026710275,
      "grad_norm": 0.09132084995508194,
      "learning_rate": 9.791153802572341e-05,
      "loss": 1.3793,
      "mean_token_accuracy": 0.7419965401291847,
      "num_tokens": 544789242.0,
      "step": 7580
    },
    {
      "entropy": 0.9550965338945389,
      "epoch": 2.0686835650040885,
      "grad_norm": 0.11602402478456497,
      "learning_rate": 9.790293687569749e-05,
      "loss": 1.3875,
      "mean_token_accuracy": 0.744350166618824,
      "num_tokens": 545518112.0,
      "step": 7590
    },
    {
      "entropy": 0.9603171899914742,
      "epoch": 2.0714091032979014,
      "grad_norm": 0.09773866832256317,
      "learning_rate": 9.789431842986284e-05,
      "loss": 1.3864,
      "mean_token_accuracy": 0.742016325891018,
      "num_tokens": 546243638.0,
      "step": 7600
    },
    {
      "entropy": 0.9539955139160157,
      "epoch": 2.0741346415917143,
      "grad_norm": 0.08773192763328552,
      "learning_rate": 9.788568269133123e-05,
      "loss": 1.3974,
      "mean_token_accuracy": 0.740584197640419,
      "num_tokens": 546965403.0,
      "step": 7610
    },
    {
      "entropy": 0.9547247931361198,
      "epoch": 2.0768601798855273,
      "grad_norm": 0.09583528339862823,
      "learning_rate": 9.78770296632207e-05,
      "loss": 1.3866,
      "mean_token_accuracy": 0.7411746963858604,
      "num_tokens": 547670412.0,
      "step": 7620
    },
    {
      "entropy": 0.9567594841122627,
      "epoch": 2.0795857181793402,
      "grad_norm": 0.11080870777368546,
      "learning_rate": 9.786835934865547e-05,
      "loss": 1.3937,
      "mean_token_accuracy": 0.7380332082509995,
      "num_tokens": 548367187.0,
      "step": 7630
    },
    {
      "entropy": 0.9504744604229927,
      "epoch": 2.0823112564731536,
      "grad_norm": 0.09894076734781265,
      "learning_rate": 9.785967175076607e-05,
      "loss": 1.3794,
      "mean_token_accuracy": 0.7484308660030365,
      "num_tokens": 549096345.0,
      "step": 7640
    },
    {
      "entropy": 0.9536508589982986,
      "epoch": 2.0850367947669666,
      "grad_norm": 0.09449561685323715,
      "learning_rate": 9.785096687268922e-05,
      "loss": 1.3866,
      "mean_token_accuracy": 0.7422194197773934,
      "num_tokens": 549813610.0,
      "step": 7650
    },
    {
      "entropy": 0.9630085244774819,
      "epoch": 2.0877623330607795,
      "grad_norm": 0.09821897745132446,
      "learning_rate": 9.784224471756791e-05,
      "loss": 1.4001,
      "mean_token_accuracy": 0.7393027573823929,
      "num_tokens": 550530533.0,
      "step": 7660
    },
    {
      "entropy": 0.9556149289011955,
      "epoch": 2.0904878713545925,
      "grad_norm": 0.09654974937438965,
      "learning_rate": 9.783350528855135e-05,
      "loss": 1.3932,
      "mean_token_accuracy": 0.7444974467158317,
      "num_tokens": 551255269.0,
      "step": 7670
    },
    {
      "entropy": 0.9613713935017586,
      "epoch": 2.0932134096484054,
      "grad_norm": 0.18773193657398224,
      "learning_rate": 9.7824748588795e-05,
      "loss": 1.3952,
      "mean_token_accuracy": 0.7414082109928131,
      "num_tokens": 551972535.0,
      "step": 7680
    },
    {
      "entropy": 0.9504840761423111,
      "epoch": 2.095938947942219,
      "grad_norm": 0.11335001140832901,
      "learning_rate": 9.781597462146055e-05,
      "loss": 1.3852,
      "mean_token_accuracy": 0.7465856090188027,
      "num_tokens": 552696476.0,
      "step": 7690
    },
    {
      "entropy": 0.9493086948990822,
      "epoch": 2.0986644862360317,
      "grad_norm": 0.10688840597867966,
      "learning_rate": 9.780718338971591e-05,
      "loss": 1.375,
      "mean_token_accuracy": 0.7483811944723129,
      "num_tokens": 553431306.0,
      "step": 7700
    },
    {
      "entropy": 0.9577390864491463,
      "epoch": 2.1013900245298447,
      "grad_norm": 0.08685535937547684,
      "learning_rate": 9.779837489673524e-05,
      "loss": 1.3937,
      "mean_token_accuracy": 0.7411364555358887,
      "num_tokens": 554157846.0,
      "step": 7710
    },
    {
      "entropy": 0.9565280213952064,
      "epoch": 2.1041155628236576,
      "grad_norm": 0.1284785121679306,
      "learning_rate": 9.778954914569896e-05,
      "loss": 1.3894,
      "mean_token_accuracy": 0.7396746635437011,
      "num_tokens": 554874935.0,
      "step": 7720
    },
    {
      "entropy": 0.9480376198887825,
      "epoch": 2.1068411011174706,
      "grad_norm": 0.09406908601522446,
      "learning_rate": 9.778070613979367e-05,
      "loss": 1.3807,
      "mean_token_accuracy": 0.7440237000584602,
      "num_tokens": 555595323.0,
      "step": 7730
    },
    {
      "entropy": 0.9471285820007325,
      "epoch": 2.1095666394112835,
      "grad_norm": 0.08738315850496292,
      "learning_rate": 9.77718458822122e-05,
      "loss": 1.3817,
      "mean_token_accuracy": 0.7418400928378105,
      "num_tokens": 556305735.0,
      "step": 7740
    },
    {
      "entropy": 0.949717079102993,
      "epoch": 2.112292177705097,
      "grad_norm": 0.11823268979787827,
      "learning_rate": 9.776296837615366e-05,
      "loss": 1.3871,
      "mean_token_accuracy": 0.7410406544804573,
      "num_tokens": 557020720.0,
      "step": 7750
    },
    {
      "entropy": 0.9500510856509209,
      "epoch": 2.11501771599891,
      "grad_norm": 0.09941519796848297,
      "learning_rate": 9.775407362482333e-05,
      "loss": 1.3845,
      "mean_token_accuracy": 0.7463141173124314,
      "num_tokens": 557741196.0,
      "step": 7760
    },
    {
      "entropy": 0.9501587927341462,
      "epoch": 2.117743254292723,
      "grad_norm": 0.10654324293136597,
      "learning_rate": 9.774516163143279e-05,
      "loss": 1.3857,
      "mean_token_accuracy": 0.7417897716164589,
      "num_tokens": 558452076.0,
      "step": 7770
    },
    {
      "entropy": 0.9522027879953384,
      "epoch": 2.1204687925865358,
      "grad_norm": 0.10403914749622345,
      "learning_rate": 9.773623239919974e-05,
      "loss": 1.3913,
      "mean_token_accuracy": 0.7401172786951065,
      "num_tokens": 559162091.0,
      "step": 7780
    },
    {
      "entropy": 0.9494891196489335,
      "epoch": 2.1231943308803487,
      "grad_norm": 0.09638772904872894,
      "learning_rate": 9.772728593134821e-05,
      "loss": 1.3753,
      "mean_token_accuracy": 0.7422588348388672,
      "num_tokens": 559888464.0,
      "step": 7790
    },
    {
      "entropy": 0.9565345466136932,
      "epoch": 2.125919869174162,
      "grad_norm": 0.09932591766119003,
      "learning_rate": 9.771832223110839e-05,
      "loss": 1.389,
      "mean_token_accuracy": 0.7491851806640625,
      "num_tokens": 560628399.0,
      "step": 7800
    },
    {
      "entropy": 0.9565155878663063,
      "epoch": 2.128645407467975,
      "grad_norm": 0.11492311209440231,
      "learning_rate": 9.77093413017167e-05,
      "loss": 1.3851,
      "mean_token_accuracy": 0.7480335250496865,
      "num_tokens": 561365171.0,
      "step": 7810
    },
    {
      "entropy": 0.9638590782880783,
      "epoch": 2.131370945761788,
      "grad_norm": 0.09567680209875107,
      "learning_rate": 9.770034314641581e-05,
      "loss": 1.3951,
      "mean_token_accuracy": 0.742455905675888,
      "num_tokens": 562086384.0,
      "step": 7820
    },
    {
      "entropy": 0.9431756198406219,
      "epoch": 2.134096484055601,
      "grad_norm": 0.09787328541278839,
      "learning_rate": 9.769132776845457e-05,
      "loss": 1.3748,
      "mean_token_accuracy": 0.7439287409186364,
      "num_tokens": 562800406.0,
      "step": 7830
    },
    {
      "entropy": 0.9539495810866356,
      "epoch": 2.136822022349414,
      "grad_norm": 0.12535595893859863,
      "learning_rate": 9.768229517108808e-05,
      "loss": 1.3917,
      "mean_token_accuracy": 0.7411072447896003,
      "num_tokens": 563509873.0,
      "step": 7840
    },
    {
      "entropy": 0.9599957779049874,
      "epoch": 2.1395475606432273,
      "grad_norm": 0.10500151664018631,
      "learning_rate": 9.767324535757765e-05,
      "loss": 1.3928,
      "mean_token_accuracy": 0.7441463723778725,
      "num_tokens": 564231965.0,
      "step": 7850
    },
    {
      "entropy": 0.9541016072034836,
      "epoch": 2.14227309893704,
      "grad_norm": 0.1107301414012909,
      "learning_rate": 9.766417833119078e-05,
      "loss": 1.3872,
      "mean_token_accuracy": 0.7417415857315064,
      "num_tokens": 564952531.0,
      "step": 7860
    },
    {
      "entropy": 0.9564834281802177,
      "epoch": 2.144998637230853,
      "grad_norm": 0.11248789727687836,
      "learning_rate": 9.765509409520122e-05,
      "loss": 1.387,
      "mean_token_accuracy": 0.7449320316314697,
      "num_tokens": 565673147.0,
      "step": 7870
    },
    {
      "entropy": 0.9606180742383004,
      "epoch": 2.147724175524666,
      "grad_norm": 0.08661828190088272,
      "learning_rate": 9.764599265288891e-05,
      "loss": 1.3914,
      "mean_token_accuracy": 0.7435157239437103,
      "num_tokens": 566393949.0,
      "step": 7880
    },
    {
      "entropy": 0.9510576590895653,
      "epoch": 2.150449713818479,
      "grad_norm": 0.09758838266134262,
      "learning_rate": 9.763687400754003e-05,
      "loss": 1.3943,
      "mean_token_accuracy": 0.7448127523064614,
      "num_tokens": 567104089.0,
      "step": 7890
    },
    {
      "entropy": 0.958026398718357,
      "epoch": 2.153175252112292,
      "grad_norm": 0.09131457656621933,
      "learning_rate": 9.762773816244694e-05,
      "loss": 1.3917,
      "mean_token_accuracy": 0.7430487379431725,
      "num_tokens": 567820065.0,
      "step": 7900
    },
    {
      "entropy": 0.9543215349316597,
      "epoch": 2.1559007904061054,
      "grad_norm": 0.11335130780935287,
      "learning_rate": 9.761858512090823e-05,
      "loss": 1.3888,
      "mean_token_accuracy": 0.7397184625267983,
      "num_tokens": 568526861.0,
      "step": 7910
    },
    {
      "entropy": 0.9609607473015785,
      "epoch": 2.1586263286999183,
      "grad_norm": 0.09940949082374573,
      "learning_rate": 9.760941488622868e-05,
      "loss": 1.3922,
      "mean_token_accuracy": 0.7399792060256004,
      "num_tokens": 569241225.0,
      "step": 7920
    },
    {
      "entropy": 0.9489357486367226,
      "epoch": 2.1613518669937313,
      "grad_norm": 0.11543278396129608,
      "learning_rate": 9.760022746171929e-05,
      "loss": 1.38,
      "mean_token_accuracy": 0.7419973850250244,
      "num_tokens": 569951679.0,
      "step": 7930
    },
    {
      "entropy": 0.9497444748878479,
      "epoch": 2.164077405287544,
      "grad_norm": 0.11987430602312088,
      "learning_rate": 9.759102285069728e-05,
      "loss": 1.3892,
      "mean_token_accuracy": 0.747117954492569,
      "num_tokens": 570678338.0,
      "step": 7940
    },
    {
      "entropy": 0.9563957571983337,
      "epoch": 2.166802943581357,
      "grad_norm": 0.1401757448911667,
      "learning_rate": 9.758180105648605e-05,
      "loss": 1.3864,
      "mean_token_accuracy": 0.7415750741958618,
      "num_tokens": 571394973.0,
      "step": 7950
    },
    {
      "entropy": 0.9495481014251709,
      "epoch": 2.1695284818751706,
      "grad_norm": 0.09080668538808823,
      "learning_rate": 9.757256208241523e-05,
      "loss": 1.3858,
      "mean_token_accuracy": 0.7434554889798164,
      "num_tokens": 572109622.0,
      "step": 7960
    },
    {
      "entropy": 0.9519291743636131,
      "epoch": 2.1722540201689835,
      "grad_norm": 0.16083116829395294,
      "learning_rate": 9.756330593182064e-05,
      "loss": 1.3884,
      "mean_token_accuracy": 0.7435447379946709,
      "num_tokens": 572809934.0,
      "step": 7970
    },
    {
      "entropy": 0.9475197449326516,
      "epoch": 2.1749795584627964,
      "grad_norm": 0.18879470229148865,
      "learning_rate": 9.755403260804427e-05,
      "loss": 1.3833,
      "mean_token_accuracy": 0.7451119691133499,
      "num_tokens": 573522270.0,
      "step": 7980
    },
    {
      "entropy": 0.9591574087738991,
      "epoch": 2.1777050967566094,
      "grad_norm": 0.09684582054615021,
      "learning_rate": 9.754474211443438e-05,
      "loss": 1.3985,
      "mean_token_accuracy": 0.7412367045879364,
      "num_tokens": 574229904.0,
      "step": 7990
    },
    {
      "entropy": 0.9530527293682098,
      "epoch": 2.1804306350504223,
      "grad_norm": 0.20667783915996552,
      "learning_rate": 9.753543445434536e-05,
      "loss": 1.3917,
      "mean_token_accuracy": 0.7400659859180451,
      "num_tokens": 574933233.0,
      "step": 8000
    },
    {
      "entropy": 0.9602208435535431,
      "epoch": 2.1831561733442353,
      "grad_norm": 0.13707610964775085,
      "learning_rate": 9.752610963113785e-05,
      "loss": 1.3933,
      "mean_token_accuracy": 0.7446024715900421,
      "num_tokens": 575657448.0,
      "step": 8010
    },
    {
      "entropy": 0.953098563849926,
      "epoch": 2.1858817116380487,
      "grad_norm": 0.09681950509548187,
      "learning_rate": 9.751676764817866e-05,
      "loss": 1.3707,
      "mean_token_accuracy": 0.7434972569346427,
      "num_tokens": 576382959.0,
      "step": 8020
    },
    {
      "entropy": 0.9519978165626526,
      "epoch": 2.1886072499318616,
      "grad_norm": 0.10283658653497696,
      "learning_rate": 9.75074085088408e-05,
      "loss": 1.386,
      "mean_token_accuracy": 0.7427999630570412,
      "num_tokens": 577099498.0,
      "step": 8030
    },
    {
      "entropy": 0.9588069602847099,
      "epoch": 2.1913327882256746,
      "grad_norm": 0.167690247297287,
      "learning_rate": 9.749803221650347e-05,
      "loss": 1.3794,
      "mean_token_accuracy": 0.7452961578965187,
      "num_tokens": 577831077.0,
      "step": 8040
    },
    {
      "entropy": 0.9571497112512588,
      "epoch": 2.1940583265194875,
      "grad_norm": 0.09857787936925888,
      "learning_rate": 9.748863877455208e-05,
      "loss": 1.3995,
      "mean_token_accuracy": 0.7413436904549598,
      "num_tokens": 578548664.0,
      "step": 8050
    },
    {
      "entropy": 0.9569678619503975,
      "epoch": 2.1967838648133005,
      "grad_norm": 0.12631528079509735,
      "learning_rate": 9.74792281863782e-05,
      "loss": 1.3868,
      "mean_token_accuracy": 0.7428177461028099,
      "num_tokens": 579269234.0,
      "step": 8060
    },
    {
      "entropy": 0.9517800152301789,
      "epoch": 2.199509403107114,
      "grad_norm": 0.11135827004909515,
      "learning_rate": 9.746980045537964e-05,
      "loss": 1.3835,
      "mean_token_accuracy": 0.7433663338422776,
      "num_tokens": 579995687.0,
      "step": 8070
    },
    {
      "entropy": 0.9407218992710114,
      "epoch": 2.202234941400927,
      "grad_norm": 0.10755167156457901,
      "learning_rate": 9.746035558496037e-05,
      "loss": 1.3792,
      "mean_token_accuracy": 0.7451389506459236,
      "num_tokens": 580707770.0,
      "step": 8080
    },
    {
      "entropy": 0.9514376983046532,
      "epoch": 2.2049604796947397,
      "grad_norm": 0.1073494628071785,
      "learning_rate": 9.745089357853052e-05,
      "loss": 1.3833,
      "mean_token_accuracy": 0.744055698812008,
      "num_tokens": 581430212.0,
      "step": 8090
    },
    {
      "entropy": 0.947685706615448,
      "epoch": 2.2076860179885527,
      "grad_norm": 0.11241379380226135,
      "learning_rate": 9.744141443950647e-05,
      "loss": 1.3827,
      "mean_token_accuracy": 0.7359423115849495,
      "num_tokens": 582131993.0,
      "step": 8100
    },
    {
      "entropy": 0.9549921721220016,
      "epoch": 2.2104115562823656,
      "grad_norm": 0.09318625926971436,
      "learning_rate": 9.743191817131072e-05,
      "loss": 1.3887,
      "mean_token_accuracy": 0.7450414016842842,
      "num_tokens": 582863912.0,
      "step": 8110
    },
    {
      "entropy": 0.9530914172530174,
      "epoch": 2.2131370945761786,
      "grad_norm": 0.12051790952682495,
      "learning_rate": 9.742240477737202e-05,
      "loss": 1.3893,
      "mean_token_accuracy": 0.7408110216259957,
      "num_tokens": 583577159.0,
      "step": 8120
    },
    {
      "entropy": 0.9535427525639534,
      "epoch": 2.215862632869992,
      "grad_norm": 0.09240037202835083,
      "learning_rate": 9.741287426112525e-05,
      "loss": 1.3882,
      "mean_token_accuracy": 0.7459387987852096,
      "num_tokens": 584307922.0,
      "step": 8130
    },
    {
      "entropy": 0.9615517109632492,
      "epoch": 2.218588171163805,
      "grad_norm": 0.10400390625,
      "learning_rate": 9.74033266260115e-05,
      "loss": 1.3839,
      "mean_token_accuracy": 0.7364679843187332,
      "num_tokens": 585015794.0,
      "step": 8140
    },
    {
      "entropy": 0.9512642607092857,
      "epoch": 2.221313709457618,
      "grad_norm": 0.1038583517074585,
      "learning_rate": 9.739376187547804e-05,
      "loss": 1.3875,
      "mean_token_accuracy": 0.7438942402601242,
      "num_tokens": 585733940.0,
      "step": 8150
    },
    {
      "entropy": 0.955040591955185,
      "epoch": 2.224039247751431,
      "grad_norm": 0.10395492613315582,
      "learning_rate": 9.738418001297831e-05,
      "loss": 1.3847,
      "mean_token_accuracy": 0.7443490579724312,
      "num_tokens": 586457898.0,
      "step": 8160
    },
    {
      "entropy": 0.9494217172265053,
      "epoch": 2.2267647860452437,
      "grad_norm": 0.11374354362487793,
      "learning_rate": 9.737458104197192e-05,
      "loss": 1.3785,
      "mean_token_accuracy": 0.7407943025231362,
      "num_tokens": 587176542.0,
      "step": 8170
    },
    {
      "entropy": 0.9520328372716904,
      "epoch": 2.229490324339057,
      "grad_norm": 0.09953740239143372,
      "learning_rate": 9.736496496592468e-05,
      "loss": 1.3818,
      "mean_token_accuracy": 0.745783768594265,
      "num_tokens": 587907654.0,
      "step": 8180
    },
    {
      "entropy": 0.9531142428517342,
      "epoch": 2.23221586263287,
      "grad_norm": 0.10314609110355377,
      "learning_rate": 9.735533178830858e-05,
      "loss": 1.3901,
      "mean_token_accuracy": 0.7458554282784462,
      "num_tokens": 588631537.0,
      "step": 8190
    },
    {
      "entropy": 0.9586337089538575,
      "epoch": 2.234941400926683,
      "grad_norm": 0.09111809730529785,
      "learning_rate": 9.734568151260175e-05,
      "loss": 1.3924,
      "mean_token_accuracy": 0.7389423623681068,
      "num_tokens": 589343536.0,
      "step": 8200
    },
    {
      "entropy": 0.9590684294700622,
      "epoch": 2.237666939220496,
      "grad_norm": 0.09666437655687332,
      "learning_rate": 9.733601414228851e-05,
      "loss": 1.3875,
      "mean_token_accuracy": 0.7413570925593376,
      "num_tokens": 590068091.0,
      "step": 8210
    },
    {
      "entropy": 0.9482115671038628,
      "epoch": 2.240392477514309,
      "grad_norm": 0.09807022660970688,
      "learning_rate": 9.732632968085936e-05,
      "loss": 1.384,
      "mean_token_accuracy": 0.7409114748239517,
      "num_tokens": 590784949.0,
      "step": 8220
    },
    {
      "entropy": 0.955726346373558,
      "epoch": 2.2431180158081223,
      "grad_norm": 0.09671087563037872,
      "learning_rate": 9.731662813181098e-05,
      "loss": 1.3935,
      "mean_token_accuracy": 0.7449436247348785,
      "num_tokens": 591515942.0,
      "step": 8230
    },
    {
      "entropy": 0.9519837841391563,
      "epoch": 2.2458435541019353,
      "grad_norm": 0.08974765986204147,
      "learning_rate": 9.730690949864619e-05,
      "loss": 1.3817,
      "mean_token_accuracy": 0.7427958875894547,
      "num_tokens": 592239233.0,
      "step": 8240
    },
    {
      "entropy": 0.9540391772985458,
      "epoch": 2.248569092395748,
      "grad_norm": 0.1783023327589035,
      "learning_rate": 9.7297173784874e-05,
      "loss": 1.3845,
      "mean_token_accuracy": 0.741515402495861,
      "num_tokens": 592951726.0,
      "step": 8250
    },
    {
      "entropy": 0.9594511583447456,
      "epoch": 2.251294630689561,
      "grad_norm": 0.10823407024145126,
      "learning_rate": 9.728742099400958e-05,
      "loss": 1.3928,
      "mean_token_accuracy": 0.7422432124614715,
      "num_tokens": 593662944.0,
      "step": 8260
    },
    {
      "entropy": 0.9567916959524154,
      "epoch": 2.254020168983374,
      "grad_norm": 0.09590677171945572,
      "learning_rate": 9.727765112957427e-05,
      "loss": 1.3883,
      "mean_token_accuracy": 0.7430309951305389,
      "num_tokens": 594390633.0,
      "step": 8270
    },
    {
      "entropy": 0.9513653293251991,
      "epoch": 2.2567457072771875,
      "grad_norm": 0.0891508013010025,
      "learning_rate": 9.726786419509559e-05,
      "loss": 1.3903,
      "mean_token_accuracy": 0.7406950652599334,
      "num_tokens": 595098083.0,
      "step": 8280
    },
    {
      "entropy": 0.9539126053452491,
      "epoch": 2.2594712455710004,
      "grad_norm": 0.11206283420324326,
      "learning_rate": 9.725806019410717e-05,
      "loss": 1.3926,
      "mean_token_accuracy": 0.744379249215126,
      "num_tokens": 595828744.0,
      "step": 8290
    },
    {
      "entropy": 0.9574821561574935,
      "epoch": 2.2621967838648134,
      "grad_norm": 0.11940894275903702,
      "learning_rate": 9.724823913014884e-05,
      "loss": 1.3832,
      "mean_token_accuracy": 0.7377721101045609,
      "num_tokens": 596546248.0,
      "step": 8300
    },
    {
      "entropy": 0.9533024594187737,
      "epoch": 2.2649223221586263,
      "grad_norm": 0.12330570816993713,
      "learning_rate": 9.723840100676662e-05,
      "loss": 1.3874,
      "mean_token_accuracy": 0.741325494647026,
      "num_tokens": 597259125.0,
      "step": 8310
    },
    {
      "entropy": 0.9586221218109131,
      "epoch": 2.2676478604524393,
      "grad_norm": 0.09389526396989822,
      "learning_rate": 9.722854582751263e-05,
      "loss": 1.3954,
      "mean_token_accuracy": 0.7408126696944237,
      "num_tokens": 597971152.0,
      "step": 8320
    },
    {
      "entropy": 0.9567618697881699,
      "epoch": 2.270373398746252,
      "grad_norm": 0.11088203638792038,
      "learning_rate": 9.72186735959452e-05,
      "loss": 1.3843,
      "mean_token_accuracy": 0.7390223950147629,
      "num_tokens": 598681678.0,
      "step": 8330
    },
    {
      "entropy": 0.9475936159491539,
      "epoch": 2.2730989370400656,
      "grad_norm": 0.10596734285354614,
      "learning_rate": 9.720878431562878e-05,
      "loss": 1.3836,
      "mean_token_accuracy": 0.7408780589699745,
      "num_tokens": 599389846.0,
      "step": 8340
    },
    {
      "entropy": 0.9543149292469024,
      "epoch": 2.2758244753338785,
      "grad_norm": 0.1165304109454155,
      "learning_rate": 9.719887799013398e-05,
      "loss": 1.3807,
      "mean_token_accuracy": 0.7443940833210945,
      "num_tokens": 600107912.0,
      "step": 8350
    },
    {
      "entropy": 0.9515037432312965,
      "epoch": 2.2785500136276915,
      "grad_norm": 0.10647770017385483,
      "learning_rate": 9.718895462303758e-05,
      "loss": 1.3774,
      "mean_token_accuracy": 0.7442094996571541,
      "num_tokens": 600830772.0,
      "step": 8360
    },
    {
      "entropy": 0.9488893762230873,
      "epoch": 2.2812755519215044,
      "grad_norm": 0.16339656710624695,
      "learning_rate": 9.71790142179225e-05,
      "loss": 1.3851,
      "mean_token_accuracy": 0.7428596884012222,
      "num_tokens": 601540728.0,
      "step": 8370
    },
    {
      "entropy": 0.9504873648285865,
      "epoch": 2.2840010902153174,
      "grad_norm": 0.10375595092773438,
      "learning_rate": 9.716905677837782e-05,
      "loss": 1.3856,
      "mean_token_accuracy": 0.74285928606987,
      "num_tokens": 602250882.0,
      "step": 8380
    },
    {
      "entropy": 0.9590310201048851,
      "epoch": 2.2867266285091308,
      "grad_norm": 0.09047789126634598,
      "learning_rate": 9.715908230799878e-05,
      "loss": 1.3882,
      "mean_token_accuracy": 0.7422109052538872,
      "num_tokens": 602977912.0,
      "step": 8390
    },
    {
      "entropy": 0.9537284761667252,
      "epoch": 2.2894521668029437,
      "grad_norm": 0.0893293023109436,
      "learning_rate": 9.714909081038675e-05,
      "loss": 1.3825,
      "mean_token_accuracy": 0.7399809136986732,
      "num_tokens": 603691403.0,
      "step": 8400
    },
    {
      "entropy": 0.9576122686266899,
      "epoch": 2.2921777050967567,
      "grad_norm": 0.09539050608873367,
      "learning_rate": 9.713908228914923e-05,
      "loss": 1.3842,
      "mean_token_accuracy": 0.7393251180648803,
      "num_tokens": 604406280.0,
      "step": 8410
    },
    {
      "entropy": 0.9496517419815064,
      "epoch": 2.2949032433905696,
      "grad_norm": 0.09857992827892303,
      "learning_rate": 9.712905674789992e-05,
      "loss": 1.3854,
      "mean_token_accuracy": 0.7475869357585907,
      "num_tokens": 605145887.0,
      "step": 8420
    },
    {
      "entropy": 0.9548046961426735,
      "epoch": 2.2976287816843826,
      "grad_norm": 0.10884623229503632,
      "learning_rate": 9.711901419025864e-05,
      "loss": 1.3828,
      "mean_token_accuracy": 0.7435204699635506,
      "num_tokens": 605872264.0,
      "step": 8430
    },
    {
      "entropy": 0.9589501142501831,
      "epoch": 2.3003543199781955,
      "grad_norm": 0.09740767627954483,
      "learning_rate": 9.710895461985133e-05,
      "loss": 1.3953,
      "mean_token_accuracy": 0.7383265286684036,
      "num_tokens": 606573504.0,
      "step": 8440
    },
    {
      "entropy": 0.9564507827162743,
      "epoch": 2.303079858272009,
      "grad_norm": 0.10725618153810501,
      "learning_rate": 9.709887804031011e-05,
      "loss": 1.4054,
      "mean_token_accuracy": 0.73534354865551,
      "num_tokens": 607266301.0,
      "step": 8450
    },
    {
      "entropy": 0.955133493244648,
      "epoch": 2.305805396565822,
      "grad_norm": 0.09258636087179184,
      "learning_rate": 9.70887844552732e-05,
      "loss": 1.3883,
      "mean_token_accuracy": 0.7400516867637634,
      "num_tokens": 607974162.0,
      "step": 8460
    },
    {
      "entropy": 0.9519719690084457,
      "epoch": 2.308530934859635,
      "grad_norm": 0.0927857905626297,
      "learning_rate": 9.707867386838502e-05,
      "loss": 1.3844,
      "mean_token_accuracy": 0.7410557478666305,
      "num_tokens": 608684509.0,
      "step": 8470
    },
    {
      "entropy": 0.9564346343278884,
      "epoch": 2.3112564731534477,
      "grad_norm": 0.10572093725204468,
      "learning_rate": 9.706854628329606e-05,
      "loss": 1.3816,
      "mean_token_accuracy": 0.7407214358448982,
      "num_tokens": 609411594.0,
      "step": 8480
    },
    {
      "entropy": 0.9542575821280479,
      "epoch": 2.3139820114472607,
      "grad_norm": 0.09347482025623322,
      "learning_rate": 9.705840170366303e-05,
      "loss": 1.3843,
      "mean_token_accuracy": 0.7433892488479614,
      "num_tokens": 610130934.0,
      "step": 8490
    },
    {
      "entropy": 0.9546264484524727,
      "epoch": 2.316707549741074,
      "grad_norm": 0.09801976382732391,
      "learning_rate": 9.704824013314866e-05,
      "loss": 1.394,
      "mean_token_accuracy": 0.7432631969451904,
      "num_tokens": 610848349.0,
      "step": 8500
    },
    {
      "entropy": 0.9565924733877182,
      "epoch": 2.319433088034887,
      "grad_norm": 0.23593464493751526,
      "learning_rate": 9.703806157542193e-05,
      "loss": 1.3905,
      "mean_token_accuracy": 0.7429133981466294,
      "num_tokens": 611561894.0,
      "step": 8510
    },
    {
      "entropy": 0.953339995443821,
      "epoch": 2.3221586263287,
      "grad_norm": 0.1044878363609314,
      "learning_rate": 9.702786603415787e-05,
      "loss": 1.386,
      "mean_token_accuracy": 0.7427751243114471,
      "num_tokens": 612278810.0,
      "step": 8520
    },
    {
      "entropy": 0.9573228195309639,
      "epoch": 2.324884164622513,
      "grad_norm": 0.11448699235916138,
      "learning_rate": 9.70176535130377e-05,
      "loss": 1.3952,
      "mean_token_accuracy": 0.7401854783296585,
      "num_tokens": 612997260.0,
      "step": 8530
    },
    {
      "entropy": 0.9552780792117119,
      "epoch": 2.327609702916326,
      "grad_norm": 0.09680186212062836,
      "learning_rate": 9.700742401574874e-05,
      "loss": 1.3954,
      "mean_token_accuracy": 0.7415038183331489,
      "num_tokens": 613710583.0,
      "step": 8540
    },
    {
      "entropy": 0.9554706335067749,
      "epoch": 2.330335241210139,
      "grad_norm": 0.1076870709657669,
      "learning_rate": 9.699717754598446e-05,
      "loss": 1.3928,
      "mean_token_accuracy": 0.742630822956562,
      "num_tokens": 614431376.0,
      "step": 8550
    },
    {
      "entropy": 0.9571814581751823,
      "epoch": 2.333060779503952,
      "grad_norm": 0.09516485035419464,
      "learning_rate": 9.698691410744442e-05,
      "loss": 1.3877,
      "mean_token_accuracy": 0.7409067109227181,
      "num_tokens": 615155180.0,
      "step": 8560
    },
    {
      "entropy": 0.955613163113594,
      "epoch": 2.335786317797765,
      "grad_norm": 0.09925924986600876,
      "learning_rate": 9.697663370383435e-05,
      "loss": 1.388,
      "mean_token_accuracy": 0.7412252381443978,
      "num_tokens": 615873451.0,
      "step": 8570
    },
    {
      "entropy": 0.9458952382206917,
      "epoch": 2.338511856091578,
      "grad_norm": 0.1026625856757164,
      "learning_rate": 9.69663363388661e-05,
      "loss": 1.3794,
      "mean_token_accuracy": 0.746001347899437,
      "num_tokens": 616604967.0,
      "step": 8580
    },
    {
      "entropy": 0.9552543684840202,
      "epoch": 2.341237394385391,
      "grad_norm": 0.1008027046918869,
      "learning_rate": 9.695602201625758e-05,
      "loss": 1.3921,
      "mean_token_accuracy": 0.7384163230657578,
      "num_tokens": 617314881.0,
      "step": 8590
    },
    {
      "entropy": 0.9545485526323318,
      "epoch": 2.343962932679204,
      "grad_norm": 0.1401461362838745,
      "learning_rate": 9.694569073973292e-05,
      "loss": 1.3885,
      "mean_token_accuracy": 0.7435328722000122,
      "num_tokens": 618041369.0,
      "step": 8600
    },
    {
      "entropy": 0.9563568353652954,
      "epoch": 2.3466884709730174,
      "grad_norm": 0.13621768355369568,
      "learning_rate": 9.693534251302232e-05,
      "loss": 1.3934,
      "mean_token_accuracy": 0.742860633134842,
      "num_tokens": 618761033.0,
      "step": 8610
    },
    {
      "entropy": 0.9503724902868271,
      "epoch": 2.3494140092668303,
      "grad_norm": 0.09177511185407639,
      "learning_rate": 9.692497733986209e-05,
      "loss": 1.3889,
      "mean_token_accuracy": 0.7402893587946892,
      "num_tokens": 619450232.0,
      "step": 8620
    },
    {
      "entropy": 0.9515706405043602,
      "epoch": 2.3521395475606433,
      "grad_norm": 0.1006256714463234,
      "learning_rate": 9.691459522399469e-05,
      "loss": 1.3818,
      "mean_token_accuracy": 0.7391259208321571,
      "num_tokens": 620154059.0,
      "step": 8630
    },
    {
      "entropy": 0.9457090362906456,
      "epoch": 2.354865085854456,
      "grad_norm": 0.10970070958137512,
      "learning_rate": 9.690419616916866e-05,
      "loss": 1.3819,
      "mean_token_accuracy": 0.7459159597754479,
      "num_tokens": 620879082.0,
      "step": 8640
    },
    {
      "entropy": 0.9572040170431138,
      "epoch": 2.357590624148269,
      "grad_norm": 0.11229894310235977,
      "learning_rate": 9.68937801791387e-05,
      "loss": 1.4004,
      "mean_token_accuracy": 0.7371783927083015,
      "num_tokens": 621574509.0,
      "step": 8650
    },
    {
      "entropy": 0.951557745039463,
      "epoch": 2.360316162442082,
      "grad_norm": 0.10168556869029999,
      "learning_rate": 9.688334725766558e-05,
      "loss": 1.3761,
      "mean_token_accuracy": 0.7517162457108497,
      "num_tokens": 622330268.0,
      "step": 8660
    },
    {
      "entropy": 0.9423753619194031,
      "epoch": 2.3630417007358955,
      "grad_norm": 0.11148504912853241,
      "learning_rate": 9.687289740851622e-05,
      "loss": 1.3792,
      "mean_token_accuracy": 0.743215948343277,
      "num_tokens": 623039876.0,
      "step": 8670
    },
    {
      "entropy": 0.9570816576480865,
      "epoch": 2.3657672390297084,
      "grad_norm": 0.11424989998340607,
      "learning_rate": 9.686243063546363e-05,
      "loss": 1.3987,
      "mean_token_accuracy": 0.7437155336141587,
      "num_tokens": 623770204.0,
      "step": 8680
    },
    {
      "entropy": 0.9582390651106835,
      "epoch": 2.3684927773235214,
      "grad_norm": 0.09436160326004028,
      "learning_rate": 9.685194694228695e-05,
      "loss": 1.3918,
      "mean_token_accuracy": 0.7428663671016693,
      "num_tokens": 624499269.0,
      "step": 8690
    },
    {
      "entropy": 0.950636301934719,
      "epoch": 2.3712183156173343,
      "grad_norm": 0.10374874621629715,
      "learning_rate": 9.68414463327714e-05,
      "loss": 1.3861,
      "mean_token_accuracy": 0.7397300824522972,
      "num_tokens": 625209851.0,
      "step": 8700
    },
    {
      "entropy": 0.9499162703752517,
      "epoch": 2.3739438539111473,
      "grad_norm": 0.09510882198810577,
      "learning_rate": 9.683092881070834e-05,
      "loss": 1.378,
      "mean_token_accuracy": 0.7441746711730957,
      "num_tokens": 625934199.0,
      "step": 8710
    },
    {
      "entropy": 0.9522706657648087,
      "epoch": 2.3766693922049607,
      "grad_norm": 0.2338542491197586,
      "learning_rate": 9.682039437989521e-05,
      "loss": 1.3807,
      "mean_token_accuracy": 0.7424178630113601,
      "num_tokens": 626660854.0,
      "step": 8720
    },
    {
      "entropy": 0.9546517834067345,
      "epoch": 2.3793949304987736,
      "grad_norm": 0.11845209449529648,
      "learning_rate": 9.680984304413556e-05,
      "loss": 1.3893,
      "mean_token_accuracy": 0.7413112193346023,
      "num_tokens": 627380960.0,
      "step": 8730
    },
    {
      "entropy": 0.9508434534072876,
      "epoch": 2.3821204687925865,
      "grad_norm": 0.15495479106903076,
      "learning_rate": 9.679927480723908e-05,
      "loss": 1.3901,
      "mean_token_accuracy": 0.7423882052302361,
      "num_tokens": 628092793.0,
      "step": 8740
    },
    {
      "entropy": 0.9567522183060646,
      "epoch": 2.3848460070863995,
      "grad_norm": 0.09430848807096481,
      "learning_rate": 9.678868967302149e-05,
      "loss": 1.3809,
      "mean_token_accuracy": 0.7474494472146034,
      "num_tokens": 628838020.0,
      "step": 8750
    },
    {
      "entropy": 0.9573053374886513,
      "epoch": 2.3875715453802124,
      "grad_norm": 0.09825961291790009,
      "learning_rate": 9.677808764530467e-05,
      "loss": 1.3882,
      "mean_token_accuracy": 0.7434219941496849,
      "num_tokens": 629564184.0,
      "step": 8760
    },
    {
      "entropy": 0.9534874022006988,
      "epoch": 2.3902970836740254,
      "grad_norm": 0.09576196223497391,
      "learning_rate": 9.67674687279166e-05,
      "loss": 1.3819,
      "mean_token_accuracy": 0.7404048323631287,
      "num_tokens": 630283422.0,
      "step": 8770
    },
    {
      "entropy": 0.9483891680836678,
      "epoch": 2.3930226219678388,
      "grad_norm": 0.19279265403747559,
      "learning_rate": 9.675683292469132e-05,
      "loss": 1.3778,
      "mean_token_accuracy": 0.7444525852799415,
      "num_tokens": 631009693.0,
      "step": 8780
    },
    {
      "entropy": 0.9520395860075951,
      "epoch": 2.3957481602616517,
      "grad_norm": 0.19635552167892456,
      "learning_rate": 9.674618023946898e-05,
      "loss": 1.3747,
      "mean_token_accuracy": 0.7479759573936462,
      "num_tokens": 631747513.0,
      "step": 8790
    },
    {
      "entropy": 0.9492999538779259,
      "epoch": 2.3984736985554647,
      "grad_norm": 0.11218266934156418,
      "learning_rate": 9.673551067609587e-05,
      "loss": 1.3832,
      "mean_token_accuracy": 0.7467947244644165,
      "num_tokens": 632472466.0,
      "step": 8800
    },
    {
      "entropy": 0.9487012833356857,
      "epoch": 2.4011992368492776,
      "grad_norm": 0.16144615411758423,
      "learning_rate": 9.672482423842429e-05,
      "loss": 1.383,
      "mean_token_accuracy": 0.7447495311498642,
      "num_tokens": 633199504.0,
      "step": 8810
    },
    {
      "entropy": 0.9547057762742043,
      "epoch": 2.403924775143091,
      "grad_norm": 0.15687504410743713,
      "learning_rate": 9.671412093031272e-05,
      "loss": 1.3849,
      "mean_token_accuracy": 0.7398101285099983,
      "num_tokens": 633907447.0,
      "step": 8820
    },
    {
      "entropy": 0.9492294996976852,
      "epoch": 2.406650313436904,
      "grad_norm": 0.18032611906528473,
      "learning_rate": 9.670340075562568e-05,
      "loss": 1.3904,
      "mean_token_accuracy": 0.7383682012557984,
      "num_tokens": 634609837.0,
      "step": 8830
    },
    {
      "entropy": 0.9546138554811477,
      "epoch": 2.409375851730717,
      "grad_norm": 0.10445669293403625,
      "learning_rate": 9.669266371823375e-05,
      "loss": 1.3869,
      "mean_token_accuracy": 0.7432780593633652,
      "num_tokens": 635331680.0,
      "step": 8840
    },
    {
      "entropy": 0.950250256061554,
      "epoch": 2.41210139002453,
      "grad_norm": 0.100125752389431,
      "learning_rate": 9.66819098220137e-05,
      "loss": 1.3895,
      "mean_token_accuracy": 0.7454440101981163,
      "num_tokens": 636041368.0,
      "step": 8850
    },
    {
      "entropy": 0.9544099509716034,
      "epoch": 2.414826928318343,
      "grad_norm": 0.08421855419874191,
      "learning_rate": 9.66711390708483e-05,
      "loss": 1.3886,
      "mean_token_accuracy": 0.7395869076251984,
      "num_tokens": 636755977.0,
      "step": 8860
    },
    {
      "entropy": 0.9561598286032676,
      "epoch": 2.4175524666121557,
      "grad_norm": 0.16309291124343872,
      "learning_rate": 9.666035146862642e-05,
      "loss": 1.3893,
      "mean_token_accuracy": 0.7409598305821419,
      "num_tokens": 637477653.0,
      "step": 8870
    },
    {
      "entropy": 0.9585184708237648,
      "epoch": 2.420278004905969,
      "grad_norm": 0.10667727142572403,
      "learning_rate": 9.664954701924304e-05,
      "loss": 1.3945,
      "mean_token_accuracy": 0.7412994131445885,
      "num_tokens": 638189167.0,
      "step": 8880
    },
    {
      "entropy": 0.9481479018926621,
      "epoch": 2.423003543199782,
      "grad_norm": 0.1399410516023636,
      "learning_rate": 9.663872572659921e-05,
      "loss": 1.3831,
      "mean_token_accuracy": 0.7428767397999764,
      "num_tokens": 638909713.0,
      "step": 8890
    },
    {
      "entropy": 0.9509454175829888,
      "epoch": 2.425729081493595,
      "grad_norm": 0.10507156699895859,
      "learning_rate": 9.662788759460205e-05,
      "loss": 1.376,
      "mean_token_accuracy": 0.7396516442298889,
      "num_tokens": 639615575.0,
      "step": 8900
    },
    {
      "entropy": 0.9558164522051811,
      "epoch": 2.428454619787408,
      "grad_norm": 0.09864579141139984,
      "learning_rate": 9.661703262716478e-05,
      "loss": 1.393,
      "mean_token_accuracy": 0.7430677577853203,
      "num_tokens": 640337164.0,
      "step": 8910
    },
    {
      "entropy": 0.9462727442383766,
      "epoch": 2.431180158081221,
      "grad_norm": 0.10743077099323273,
      "learning_rate": 9.660616082820669e-05,
      "loss": 1.3807,
      "mean_token_accuracy": 0.7418262407183647,
      "num_tokens": 641051774.0,
      "step": 8920
    },
    {
      "entropy": 0.9490568324923515,
      "epoch": 2.4339056963750343,
      "grad_norm": 0.5460764169692993,
      "learning_rate": 9.659527220165312e-05,
      "loss": 1.3784,
      "mean_token_accuracy": 0.7486709088087082,
      "num_tokens": 641785779.0,
      "step": 8930
    },
    {
      "entropy": 0.9565368533134461,
      "epoch": 2.4366312346688472,
      "grad_norm": 0.1278989315032959,
      "learning_rate": 9.658436675143555e-05,
      "loss": 1.3889,
      "mean_token_accuracy": 0.7437928393483162,
      "num_tokens": 642509140.0,
      "step": 8940
    },
    {
      "entropy": 0.9486685395240784,
      "epoch": 2.43935677296266,
      "grad_norm": 0.1757955104112625,
      "learning_rate": 9.657344448149146e-05,
      "loss": 1.3801,
      "mean_token_accuracy": 0.7438680246472359,
      "num_tokens": 643235408.0,
      "step": 8950
    },
    {
      "entropy": 0.9519594490528107,
      "epoch": 2.442082311256473,
      "grad_norm": 0.11428218334913254,
      "learning_rate": 9.656250539576445e-05,
      "loss": 1.3898,
      "mean_token_accuracy": 0.742796714603901,
      "num_tokens": 643955009.0,
      "step": 8960
    },
    {
      "entropy": 0.948516134917736,
      "epoch": 2.444807849550286,
      "grad_norm": 0.10164281725883484,
      "learning_rate": 9.655154949820421e-05,
      "loss": 1.3788,
      "mean_token_accuracy": 0.7450691372156143,
      "num_tokens": 644683818.0,
      "step": 8970
    },
    {
      "entropy": 0.9490355685353279,
      "epoch": 2.447533387844099,
      "grad_norm": 0.10648229718208313,
      "learning_rate": 9.654057679276644e-05,
      "loss": 1.3777,
      "mean_token_accuracy": 0.746142016351223,
      "num_tokens": 645424535.0,
      "step": 8980
    },
    {
      "entropy": 0.9564800366759301,
      "epoch": 2.4502589261379124,
      "grad_norm": 0.12165772169828415,
      "learning_rate": 9.652958728341296e-05,
      "loss": 1.3956,
      "mean_token_accuracy": 0.7393794029951095,
      "num_tokens": 646132176.0,
      "step": 8990
    },
    {
      "entropy": 0.956525768339634,
      "epoch": 2.4529844644317254,
      "grad_norm": 0.14618892967700958,
      "learning_rate": 9.65185809741116e-05,
      "loss": 1.3853,
      "mean_token_accuracy": 0.743808300793171,
      "num_tokens": 646859208.0,
      "step": 9000
    },
    {
      "entropy": 0.9516721293330193,
      "epoch": 2.4557100027255383,
      "grad_norm": 0.09927177429199219,
      "learning_rate": 9.650755786883633e-05,
      "loss": 1.3814,
      "mean_token_accuracy": 0.7426035419106484,
      "num_tokens": 647585717.0,
      "step": 9010
    },
    {
      "entropy": 0.9528895154595375,
      "epoch": 2.4584355410193512,
      "grad_norm": 0.1371716558933258,
      "learning_rate": 9.649651797156713e-05,
      "loss": 1.3915,
      "mean_token_accuracy": 0.7415006205439567,
      "num_tokens": 648299102.0,
      "step": 9020
    },
    {
      "entropy": 0.9539487212896347,
      "epoch": 2.461161079313164,
      "grad_norm": 0.12393172085285187,
      "learning_rate": 9.648546128629006e-05,
      "loss": 1.3872,
      "mean_token_accuracy": 0.7432716667652131,
      "num_tokens": 649014355.0,
      "step": 9030
    },
    {
      "entropy": 0.9543512776494026,
      "epoch": 2.4638866176069776,
      "grad_norm": 0.0934886708855629,
      "learning_rate": 9.647438781699725e-05,
      "loss": 1.3817,
      "mean_token_accuracy": 0.7445197448134422,
      "num_tokens": 649738408.0,
      "step": 9040
    },
    {
      "entropy": 0.9564831212162972,
      "epoch": 2.4666121559007905,
      "grad_norm": 0.12876413762569427,
      "learning_rate": 9.646329756768686e-05,
      "loss": 1.3919,
      "mean_token_accuracy": 0.7444147065281868,
      "num_tokens": 650460693.0,
      "step": 9050
    },
    {
      "entropy": 0.95407083183527,
      "epoch": 2.4693376941946035,
      "grad_norm": 0.42325589060783386,
      "learning_rate": 9.645219054236314e-05,
      "loss": 1.3921,
      "mean_token_accuracy": 0.7388138264417649,
      "num_tokens": 651171240.0,
      "step": 9060
    },
    {
      "entropy": 0.9518125399947166,
      "epoch": 2.4720632324884164,
      "grad_norm": 0.10088993608951569,
      "learning_rate": 9.644106674503638e-05,
      "loss": 1.3839,
      "mean_token_accuracy": 0.7464907243847847,
      "num_tokens": 651901415.0,
      "step": 9070
    },
    {
      "entropy": 0.9502801403403283,
      "epoch": 2.4747887707822294,
      "grad_norm": 0.09579727053642273,
      "learning_rate": 9.642992617972297e-05,
      "loss": 1.3814,
      "mean_token_accuracy": 0.7473923444747925,
      "num_tokens": 652632701.0,
      "step": 9080
    },
    {
      "entropy": 0.9594358250498771,
      "epoch": 2.4775143090760423,
      "grad_norm": 0.08653602004051208,
      "learning_rate": 9.641876885044528e-05,
      "loss": 1.3939,
      "mean_token_accuracy": 0.7395642042160034,
      "num_tokens": 653352424.0,
      "step": 9090
    },
    {
      "entropy": 0.9477410703897476,
      "epoch": 2.4802398473698557,
      "grad_norm": 0.09166956692934036,
      "learning_rate": 9.640759476123176e-05,
      "loss": 1.3947,
      "mean_token_accuracy": 0.7414721310138702,
      "num_tokens": 654058073.0,
      "step": 9100
    },
    {
      "entropy": 0.952572925388813,
      "epoch": 2.4829653856636686,
      "grad_norm": 0.10699083656072617,
      "learning_rate": 9.639640391611695e-05,
      "loss": 1.3882,
      "mean_token_accuracy": 0.7433302894234657,
      "num_tokens": 654770000.0,
      "step": 9110
    },
    {
      "entropy": 0.9538148298859597,
      "epoch": 2.4856909239574816,
      "grad_norm": 0.132701113820076,
      "learning_rate": 9.63851963191414e-05,
      "loss": 1.3892,
      "mean_token_accuracy": 0.7426132261753082,
      "num_tokens": 655492183.0,
      "step": 9120
    },
    {
      "entropy": 0.9542162343859673,
      "epoch": 2.4884164622512945,
      "grad_norm": 0.12824146449565887,
      "learning_rate": 9.63739719743517e-05,
      "loss": 1.3855,
      "mean_token_accuracy": 0.742797639966011,
      "num_tokens": 656205921.0,
      "step": 9130
    },
    {
      "entropy": 0.9498761922121048,
      "epoch": 2.4911420005451075,
      "grad_norm": 0.093867227435112,
      "learning_rate": 9.636273088580054e-05,
      "loss": 1.3866,
      "mean_token_accuracy": 0.7396522730588913,
      "num_tokens": 656912518.0,
      "step": 9140
    },
    {
      "entropy": 0.9562533587217331,
      "epoch": 2.493867538838921,
      "grad_norm": 0.15160538256168365,
      "learning_rate": 9.63514730575466e-05,
      "loss": 1.3845,
      "mean_token_accuracy": 0.7418175086379051,
      "num_tokens": 657624221.0,
      "step": 9150
    },
    {
      "entropy": 0.9461548939347267,
      "epoch": 2.496593077132734,
      "grad_norm": 0.09553908556699753,
      "learning_rate": 9.634019849365463e-05,
      "loss": 1.3816,
      "mean_token_accuracy": 0.7410569325089454,
      "num_tokens": 658335159.0,
      "step": 9160
    },
    {
      "entropy": 0.9439819872379303,
      "epoch": 2.4993186154265468,
      "grad_norm": 0.1621396392583847,
      "learning_rate": 9.632890719819543e-05,
      "loss": 1.3771,
      "mean_token_accuracy": 0.7506534278392791,
      "num_tokens": 659067134.0,
      "step": 9170
    },
    {
      "entropy": 0.9528698980808258,
      "epoch": 2.5020441537203597,
      "grad_norm": 0.09267479181289673,
      "learning_rate": 9.631759917524582e-05,
      "loss": 1.3845,
      "mean_token_accuracy": 0.7421308100223541,
      "num_tokens": 659777107.0,
      "step": 9180
    },
    {
      "entropy": 0.9538725689053535,
      "epoch": 2.5047696920141727,
      "grad_norm": 0.09763748198747635,
      "learning_rate": 9.630627442888864e-05,
      "loss": 1.3961,
      "mean_token_accuracy": 0.7397515043616295,
      "num_tokens": 660488201.0,
      "step": 9190
    },
    {
      "entropy": 0.9516035065054893,
      "epoch": 2.5074952303079856,
      "grad_norm": 0.08752143383026123,
      "learning_rate": 9.629493296321284e-05,
      "loss": 1.3774,
      "mean_token_accuracy": 0.7428139641880989,
      "num_tokens": 661208951.0,
      "step": 9200
    },
    {
      "entropy": 0.958625254034996,
      "epoch": 2.510220768601799,
      "grad_norm": 0.11285807937383652,
      "learning_rate": 9.628357478231334e-05,
      "loss": 1.3825,
      "mean_token_accuracy": 0.7390116930007935,
      "num_tokens": 661922597.0,
      "step": 9210
    },
    {
      "entropy": 0.9498637214303016,
      "epoch": 2.512946306895612,
      "grad_norm": 0.10862982273101807,
      "learning_rate": 9.627219989029113e-05,
      "loss": 1.3894,
      "mean_token_accuracy": 0.7430294454097748,
      "num_tokens": 662631760.0,
      "step": 9220
    },
    {
      "entropy": 0.9558670163154602,
      "epoch": 2.515671845189425,
      "grad_norm": 0.09293710440397263,
      "learning_rate": 9.62608082912532e-05,
      "loss": 1.3845,
      "mean_token_accuracy": 0.7393468782305718,
      "num_tokens": 663334315.0,
      "step": 9230
    },
    {
      "entropy": 0.9459223300218582,
      "epoch": 2.518397383483238,
      "grad_norm": 0.11126852035522461,
      "learning_rate": 9.624939998931263e-05,
      "loss": 1.3843,
      "mean_token_accuracy": 0.7441275328397751,
      "num_tokens": 664053780.0,
      "step": 9240
    },
    {
      "entropy": 0.948776651918888,
      "epoch": 2.521122921777051,
      "grad_norm": 0.1200125515460968,
      "learning_rate": 9.623797498858846e-05,
      "loss": 1.3753,
      "mean_token_accuracy": 0.7464207157492637,
      "num_tokens": 664789320.0,
      "step": 9250
    },
    {
      "entropy": 0.9534156277775765,
      "epoch": 2.523848460070864,
      "grad_norm": 0.10603691637516022,
      "learning_rate": 9.622653329320583e-05,
      "loss": 1.383,
      "mean_token_accuracy": 0.7419575944542884,
      "num_tokens": 665512089.0,
      "step": 9260
    },
    {
      "entropy": 0.950544461607933,
      "epoch": 2.526573998364677,
      "grad_norm": 0.09827824681997299,
      "learning_rate": 9.621507490729585e-05,
      "loss": 1.3873,
      "mean_token_accuracy": 0.7419786721467971,
      "num_tokens": 666223704.0,
      "step": 9270
    },
    {
      "entropy": 0.9486217007040978,
      "epoch": 2.52929953665849,
      "grad_norm": 0.15562506020069122,
      "learning_rate": 9.620359983499568e-05,
      "loss": 1.3748,
      "mean_token_accuracy": 0.7466563552618026,
      "num_tokens": 666953383.0,
      "step": 9280
    },
    {
      "entropy": 0.9574059993028641,
      "epoch": 2.532025074952303,
      "grad_norm": 0.1290716677904129,
      "learning_rate": 9.619210808044851e-05,
      "loss": 1.3924,
      "mean_token_accuracy": 0.7437079384922981,
      "num_tokens": 667666243.0,
      "step": 9290
    },
    {
      "entropy": 0.9493744015693665,
      "epoch": 2.534750613246116,
      "grad_norm": 0.09293576329946518,
      "learning_rate": 9.618059964780354e-05,
      "loss": 1.3787,
      "mean_token_accuracy": 0.7413115292787552,
      "num_tokens": 668377955.0,
      "step": 9300
    },
    {
      "entropy": 0.9519496336579323,
      "epoch": 2.537476151539929,
      "grad_norm": 0.1295005977153778,
      "learning_rate": 9.6169074541216e-05,
      "loss": 1.3893,
      "mean_token_accuracy": 0.7420684978365898,
      "num_tokens": 669095589.0,
      "step": 9310
    },
    {
      "entropy": 0.9614164233207703,
      "epoch": 2.5402016898337423,
      "grad_norm": 0.10095058381557465,
      "learning_rate": 9.615753276484715e-05,
      "loss": 1.3885,
      "mean_token_accuracy": 0.7394142135977745,
      "num_tokens": 669800172.0,
      "step": 9320
    },
    {
      "entropy": 0.9443493977189064,
      "epoch": 2.5429272281275552,
      "grad_norm": 0.09708225727081299,
      "learning_rate": 9.614597432286425e-05,
      "loss": 1.3768,
      "mean_token_accuracy": 0.7487831532955169,
      "num_tokens": 670532131.0,
      "step": 9330
    },
    {
      "entropy": 0.954305274784565,
      "epoch": 2.545652766421368,
      "grad_norm": 0.08251254260540009,
      "learning_rate": 9.613439921944056e-05,
      "loss": 1.395,
      "mean_token_accuracy": 0.7416146233677864,
      "num_tokens": 671243931.0,
      "step": 9340
    },
    {
      "entropy": 0.9444291368126869,
      "epoch": 2.548378304715181,
      "grad_norm": 0.1446608603000641,
      "learning_rate": 9.612280745875543e-05,
      "loss": 1.3709,
      "mean_token_accuracy": 0.7403537198901177,
      "num_tokens": 671953439.0,
      "step": 9350
    },
    {
      "entropy": 0.9425195306539536,
      "epoch": 2.5511038430089945,
      "grad_norm": 0.11479167640209198,
      "learning_rate": 9.611119904499413e-05,
      "loss": 1.3803,
      "mean_token_accuracy": 0.744185908138752,
      "num_tokens": 672668357.0,
      "step": 9360
    },
    {
      "entropy": 0.9567700877785683,
      "epoch": 2.5538293813028075,
      "grad_norm": 0.09406740963459015,
      "learning_rate": 9.609957398234801e-05,
      "loss": 1.3844,
      "mean_token_accuracy": 0.7394072458148002,
      "num_tokens": 673379901.0,
      "step": 9370
    },
    {
      "entropy": 0.9458857789635658,
      "epoch": 2.5565549195966204,
      "grad_norm": 0.12724296748638153,
      "learning_rate": 9.608793227501442e-05,
      "loss": 1.3831,
      "mean_token_accuracy": 0.7463359922170639,
      "num_tokens": 674105083.0,
      "step": 9380
    },
    {
      "entropy": 0.9527260407805442,
      "epoch": 2.5592804578904333,
      "grad_norm": 0.08779218792915344,
      "learning_rate": 9.607627392719669e-05,
      "loss": 1.3894,
      "mean_token_accuracy": 0.7400561019778251,
      "num_tokens": 674812501.0,
      "step": 9390
    },
    {
      "entropy": 0.9508476212620736,
      "epoch": 2.5620059961842463,
      "grad_norm": 0.11054495722055435,
      "learning_rate": 9.606459894310416e-05,
      "loss": 1.3747,
      "mean_token_accuracy": 0.7446539625525475,
      "num_tokens": 675537383.0,
      "step": 9400
    },
    {
      "entropy": 0.9514044508337974,
      "epoch": 2.5647315344780592,
      "grad_norm": 0.10721174627542496,
      "learning_rate": 9.605290732695223e-05,
      "loss": 1.3841,
      "mean_token_accuracy": 0.7457584992051125,
      "num_tokens": 676269799.0,
      "step": 9410
    },
    {
      "entropy": 0.9455446422100067,
      "epoch": 2.567457072771872,
      "grad_norm": 0.13902905583381653,
      "learning_rate": 9.604119908296226e-05,
      "loss": 1.3803,
      "mean_token_accuracy": 0.7467570066452026,
      "num_tokens": 676984727.0,
      "step": 9420
    },
    {
      "entropy": 0.9542788952589035,
      "epoch": 2.5701826110656856,
      "grad_norm": 0.09337563067674637,
      "learning_rate": 9.60294742153616e-05,
      "loss": 1.377,
      "mean_token_accuracy": 0.7402194425463676,
      "num_tokens": 677701384.0,
      "step": 9430
    },
    {
      "entropy": 0.9549151852726936,
      "epoch": 2.5729081493594985,
      "grad_norm": 0.09679771959781647,
      "learning_rate": 9.601773272838364e-05,
      "loss": 1.3864,
      "mean_token_accuracy": 0.744145430624485,
      "num_tokens": 678428010.0,
      "step": 9440
    },
    {
      "entropy": 0.9494045555591584,
      "epoch": 2.5756336876533115,
      "grad_norm": 0.09665834158658981,
      "learning_rate": 9.600597462626775e-05,
      "loss": 1.3814,
      "mean_token_accuracy": 0.7428949892520904,
      "num_tokens": 679150837.0,
      "step": 9450
    },
    {
      "entropy": 0.9528696566820145,
      "epoch": 2.5783592259471244,
      "grad_norm": 0.12176550924777985,
      "learning_rate": 9.599419991325933e-05,
      "loss": 1.3822,
      "mean_token_accuracy": 0.7424042224884033,
      "num_tokens": 679871972.0,
      "step": 9460
    },
    {
      "entropy": 0.9489214882254601,
      "epoch": 2.581084764240938,
      "grad_norm": 0.1010860800743103,
      "learning_rate": 9.598240859360972e-05,
      "loss": 1.3821,
      "mean_token_accuracy": 0.7403257504105568,
      "num_tokens": 680571890.0,
      "step": 9470
    },
    {
      "entropy": 0.949827016890049,
      "epoch": 2.5838103025347507,
      "grad_norm": 0.08193811774253845,
      "learning_rate": 9.597060067157629e-05,
      "loss": 1.3939,
      "mean_token_accuracy": 0.7420325294137001,
      "num_tokens": 681290271.0,
      "step": 9480
    },
    {
      "entropy": 0.9546469509601593,
      "epoch": 2.5865358408285637,
      "grad_norm": 0.1042378768324852,
      "learning_rate": 9.595877615142241e-05,
      "loss": 1.3878,
      "mean_token_accuracy": 0.7438606515526771,
      "num_tokens": 682005765.0,
      "step": 9490
    },
    {
      "entropy": 0.9530891895294189,
      "epoch": 2.5892613791223766,
      "grad_norm": 0.09580600261688232,
      "learning_rate": 9.594693503741745e-05,
      "loss": 1.3861,
      "mean_token_accuracy": 0.7394726812839508,
      "num_tokens": 682706989.0,
      "step": 9500
    },
    {
      "entropy": 0.9467429131269455,
      "epoch": 2.5919869174161896,
      "grad_norm": 0.08970653265714645,
      "learning_rate": 9.593507733383673e-05,
      "loss": 1.3879,
      "mean_token_accuracy": 0.7434066340327263,
      "num_tokens": 683420796.0,
      "step": 9510
    },
    {
      "entropy": 0.9422464728355407,
      "epoch": 2.5947124557100025,
      "grad_norm": 0.10844879597425461,
      "learning_rate": 9.592320304496159e-05,
      "loss": 1.3745,
      "mean_token_accuracy": 0.7470636069774628,
      "num_tokens": 684155827.0,
      "step": 9520
    },
    {
      "entropy": 0.9513745456933975,
      "epoch": 2.597437994003816,
      "grad_norm": 0.09567851573228836,
      "learning_rate": 9.591131217507938e-05,
      "loss": 1.385,
      "mean_token_accuracy": 0.7429541811347008,
      "num_tokens": 684872588.0,
      "step": 9530
    },
    {
      "entropy": 0.9407967448234558,
      "epoch": 2.600163532297629,
      "grad_norm": 0.11687707901000977,
      "learning_rate": 9.589940472848338e-05,
      "loss": 1.3779,
      "mean_token_accuracy": 0.7470522150397301,
      "num_tokens": 685600844.0,
      "step": 9540
    },
    {
      "entropy": 0.9530697196722031,
      "epoch": 2.602889070591442,
      "grad_norm": 0.10524759441614151,
      "learning_rate": 9.588748070947289e-05,
      "loss": 1.3895,
      "mean_token_accuracy": 0.7454262211918831,
      "num_tokens": 686340335.0,
      "step": 9550
    },
    {
      "entropy": 0.9604225113987923,
      "epoch": 2.6056146088852548,
      "grad_norm": 0.1046706959605217,
      "learning_rate": 9.58755401223532e-05,
      "loss": 1.3918,
      "mean_token_accuracy": 0.7387133181095124,
      "num_tokens": 687044113.0,
      "step": 9560
    },
    {
      "entropy": 0.9585466727614402,
      "epoch": 2.6083401471790677,
      "grad_norm": 0.09597111493349075,
      "learning_rate": 9.586358297143558e-05,
      "loss": 1.3922,
      "mean_token_accuracy": 0.7404394149780273,
      "num_tokens": 687757839.0,
      "step": 9570
    },
    {
      "entropy": 0.9465219035744667,
      "epoch": 2.611065685472881,
      "grad_norm": 0.368947833776474,
      "learning_rate": 9.585160926103727e-05,
      "loss": 1.3909,
      "mean_token_accuracy": 0.7451237291097641,
      "num_tokens": 688480525.0,
      "step": 9580
    },
    {
      "entropy": 0.9518843427300453,
      "epoch": 2.613791223766694,
      "grad_norm": 0.09533551335334778,
      "learning_rate": 9.583961899548149e-05,
      "loss": 1.377,
      "mean_token_accuracy": 0.7444559141993523,
      "num_tokens": 689211956.0,
      "step": 9590
    },
    {
      "entropy": 0.9531243652105331,
      "epoch": 2.616516762060507,
      "grad_norm": 0.18116629123687744,
      "learning_rate": 9.582761217909742e-05,
      "loss": 1.3839,
      "mean_token_accuracy": 0.7390112906694413,
      "num_tokens": 689925357.0,
      "step": 9600
    },
    {
      "entropy": 0.9519005209207535,
      "epoch": 2.61924230035432,
      "grad_norm": 0.11106158792972565,
      "learning_rate": 9.581558881622023e-05,
      "loss": 1.3803,
      "mean_token_accuracy": 0.7452313512563705,
      "num_tokens": 690648163.0,
      "step": 9610
    },
    {
      "entropy": 0.9523310184478759,
      "epoch": 2.621967838648133,
      "grad_norm": 0.11411701887845993,
      "learning_rate": 9.580354891119111e-05,
      "loss": 1.3932,
      "mean_token_accuracy": 0.741227887570858,
      "num_tokens": 691368788.0,
      "step": 9620
    },
    {
      "entropy": 0.9491930544376374,
      "epoch": 2.624693376941946,
      "grad_norm": 0.10222488641738892,
      "learning_rate": 9.579149246835714e-05,
      "loss": 1.3753,
      "mean_token_accuracy": 0.744267588853836,
      "num_tokens": 692098494.0,
      "step": 9630
    },
    {
      "entropy": 0.9463312938809395,
      "epoch": 2.627418915235759,
      "grad_norm": 0.10576868802309036,
      "learning_rate": 9.577941949207146e-05,
      "loss": 1.3765,
      "mean_token_accuracy": 0.7458298191428184,
      "num_tokens": 692828764.0,
      "step": 9640
    },
    {
      "entropy": 0.9565997913479805,
      "epoch": 2.630144453529572,
      "grad_norm": 0.10648185014724731,
      "learning_rate": 9.576732998669309e-05,
      "loss": 1.3922,
      "mean_token_accuracy": 0.7440996661782264,
      "num_tokens": 693536417.0,
      "step": 9650
    },
    {
      "entropy": 0.952058082818985,
      "epoch": 2.632869991823385,
      "grad_norm": 0.09439050406217575,
      "learning_rate": 9.575522395658707e-05,
      "loss": 1.3861,
      "mean_token_accuracy": 0.7394286692142487,
      "num_tokens": 694243626.0,
      "step": 9660
    },
    {
      "entropy": 0.953196357190609,
      "epoch": 2.635595530117198,
      "grad_norm": 0.11707699298858643,
      "learning_rate": 9.574310140612442e-05,
      "loss": 1.3827,
      "mean_token_accuracy": 0.7434344187378883,
      "num_tokens": 694968647.0,
      "step": 9670
    },
    {
      "entropy": 0.939579950273037,
      "epoch": 2.6383210684110114,
      "grad_norm": 0.2158738225698471,
      "learning_rate": 9.573096233968206e-05,
      "loss": 1.3704,
      "mean_token_accuracy": 0.7507406219840049,
      "num_tokens": 695704102.0,
      "step": 9680
    },
    {
      "entropy": 0.9536104574799538,
      "epoch": 2.6410466067048244,
      "grad_norm": 0.13156963884830475,
      "learning_rate": 9.571880676164293e-05,
      "loss": 1.3835,
      "mean_token_accuracy": 0.7448409900069237,
      "num_tokens": 696426806.0,
      "step": 9690
    },
    {
      "entropy": 0.9547406882047653,
      "epoch": 2.6437721449986373,
      "grad_norm": 0.09959427267313004,
      "learning_rate": 9.570663467639595e-05,
      "loss": 1.384,
      "mean_token_accuracy": 0.7404066026210785,
      "num_tokens": 697150365.0,
      "step": 9700
    },
    {
      "entropy": 0.9502587810158729,
      "epoch": 2.6464976832924503,
      "grad_norm": 0.0893438532948494,
      "learning_rate": 9.569444608833593e-05,
      "loss": 1.3832,
      "mean_token_accuracy": 0.7417907059192658,
      "num_tokens": 697864829.0,
      "step": 9710
    },
    {
      "entropy": 0.9574382051825523,
      "epoch": 2.649223221586263,
      "grad_norm": 0.12412620335817337,
      "learning_rate": 9.568224100186368e-05,
      "loss": 1.3926,
      "mean_token_accuracy": 0.7443241059780121,
      "num_tokens": 698593171.0,
      "step": 9720
    },
    {
      "entropy": 0.9479583531618119,
      "epoch": 2.651948759880076,
      "grad_norm": 0.09614630043506622,
      "learning_rate": 9.567001942138596e-05,
      "loss": 1.377,
      "mean_token_accuracy": 0.7478373780846596,
      "num_tokens": 699328705.0,
      "step": 9730
    },
    {
      "entropy": 0.9494291916489601,
      "epoch": 2.654674298173889,
      "grad_norm": 0.21806296706199646,
      "learning_rate": 9.565778135131551e-05,
      "loss": 1.3889,
      "mean_token_accuracy": 0.7439940094947814,
      "num_tokens": 700046182.0,
      "step": 9740
    },
    {
      "entropy": 0.9508288979530335,
      "epoch": 2.6573998364677025,
      "grad_norm": 0.09569336473941803,
      "learning_rate": 9.5645526796071e-05,
      "loss": 1.3844,
      "mean_token_accuracy": 0.7418312415480613,
      "num_tokens": 700756359.0,
      "step": 9750
    },
    {
      "entropy": 0.9505010828375816,
      "epoch": 2.6601253747615154,
      "grad_norm": 0.0998583659529686,
      "learning_rate": 9.563325576007701e-05,
      "loss": 1.379,
      "mean_token_accuracy": 0.7458802133798599,
      "num_tokens": 701477494.0,
      "step": 9760
    },
    {
      "entropy": 0.9552866011857987,
      "epoch": 2.6628509130553284,
      "grad_norm": 0.14993654191493988,
      "learning_rate": 9.562096824776416e-05,
      "loss": 1.3876,
      "mean_token_accuracy": 0.7408042192459107,
      "num_tokens": 702194293.0,
      "step": 9770
    },
    {
      "entropy": 0.9506648227572441,
      "epoch": 2.6655764513491413,
      "grad_norm": 0.09288034588098526,
      "learning_rate": 9.560866426356894e-05,
      "loss": 1.3854,
      "mean_token_accuracy": 0.7435394883155823,
      "num_tokens": 702911966.0,
      "step": 9780
    },
    {
      "entropy": 0.9530773386359215,
      "epoch": 2.6683019896429547,
      "grad_norm": 0.1780327409505844,
      "learning_rate": 9.559634381193385e-05,
      "loss": 1.384,
      "mean_token_accuracy": 0.7413442313671113,
      "num_tokens": 703621324.0,
      "step": 9790
    },
    {
      "entropy": 0.9432814002037049,
      "epoch": 2.6710275279367677,
      "grad_norm": 0.08851876109838486,
      "learning_rate": 9.558400689730728e-05,
      "loss": 1.3716,
      "mean_token_accuracy": 0.7464269071817398,
      "num_tokens": 704353647.0,
      "step": 9800
    },
    {
      "entropy": 0.950922329723835,
      "epoch": 2.6737530662305806,
      "grad_norm": 0.08739480376243591,
      "learning_rate": 9.557165352414361e-05,
      "loss": 1.3777,
      "mean_token_accuracy": 0.7431806236505508,
      "num_tokens": 705083830.0,
      "step": 9810
    },
    {
      "entropy": 0.9586237475275994,
      "epoch": 2.6764786045243936,
      "grad_norm": 0.11277187615633011,
      "learning_rate": 9.555928369690313e-05,
      "loss": 1.3908,
      "mean_token_accuracy": 0.7428459405899048,
      "num_tokens": 705800066.0,
      "step": 9820
    },
    {
      "entropy": 0.9557994589209556,
      "epoch": 2.6792041428182065,
      "grad_norm": 0.08675989508628845,
      "learning_rate": 9.554689742005209e-05,
      "loss": 1.3898,
      "mean_token_accuracy": 0.7398327201604843,
      "num_tokens": 706506042.0,
      "step": 9830
    },
    {
      "entropy": 0.94582679271698,
      "epoch": 2.6819296811120195,
      "grad_norm": 0.2761112153530121,
      "learning_rate": 9.553449469806266e-05,
      "loss": 1.3735,
      "mean_token_accuracy": 0.744458457827568,
      "num_tokens": 707234275.0,
      "step": 9840
    },
    {
      "entropy": 0.9533609345555305,
      "epoch": 2.6846552194058324,
      "grad_norm": 0.15056166052818298,
      "learning_rate": 9.552207553541296e-05,
      "loss": 1.3905,
      "mean_token_accuracy": 0.7413505062460899,
      "num_tokens": 707939097.0,
      "step": 9850
    },
    {
      "entropy": 0.9512269735336304,
      "epoch": 2.687380757699646,
      "grad_norm": 0.10775098949670792,
      "learning_rate": 9.550963993658708e-05,
      "loss": 1.3877,
      "mean_token_accuracy": 0.7441338151693344,
      "num_tokens": 708662822.0,
      "step": 9860
    },
    {
      "entropy": 0.9576830640435219,
      "epoch": 2.6901062959934587,
      "grad_norm": 0.10758344829082489,
      "learning_rate": 9.549718790607497e-05,
      "loss": 1.3862,
      "mean_token_accuracy": 0.7405988663434983,
      "num_tokens": 709385510.0,
      "step": 9870
    },
    {
      "entropy": 0.9462365791201591,
      "epoch": 2.6928318342872717,
      "grad_norm": 0.16311097145080566,
      "learning_rate": 9.548471944837258e-05,
      "loss": 1.3817,
      "mean_token_accuracy": 0.7423205584287643,
      "num_tokens": 710088991.0,
      "step": 9880
    },
    {
      "entropy": 0.9475960016250611,
      "epoch": 2.6955573725810846,
      "grad_norm": 0.09887238591909409,
      "learning_rate": 9.547223456798175e-05,
      "loss": 1.3742,
      "mean_token_accuracy": 0.7462312802672386,
      "num_tokens": 710812777.0,
      "step": 9890
    },
    {
      "entropy": 0.9553654506802559,
      "epoch": 2.698282910874898,
      "grad_norm": 0.14105680584907532,
      "learning_rate": 9.545973326941026e-05,
      "loss": 1.3856,
      "mean_token_accuracy": 0.7394596874713898,
      "num_tokens": 711522164.0,
      "step": 9900
    },
    {
      "entropy": 0.9541070491075516,
      "epoch": 2.701008449168711,
      "grad_norm": 0.10180029273033142,
      "learning_rate": 9.544721555717183e-05,
      "loss": 1.3819,
      "mean_token_accuracy": 0.7387867987155914,
      "num_tokens": 712232297.0,
      "step": 9910
    },
    {
      "entropy": 0.9523576036095619,
      "epoch": 2.703733987462524,
      "grad_norm": 0.11639035493135452,
      "learning_rate": 9.543468143578612e-05,
      "loss": 1.3852,
      "mean_token_accuracy": 0.741235437989235,
      "num_tokens": 712955332.0,
      "step": 9920
    },
    {
      "entropy": 0.9462679713964463,
      "epoch": 2.706459525756337,
      "grad_norm": 0.11618001013994217,
      "learning_rate": 9.542213090977865e-05,
      "loss": 1.381,
      "mean_token_accuracy": 0.7462039306759835,
      "num_tokens": 713679744.0,
      "step": 9930
    },
    {
      "entropy": 0.9466747432947159,
      "epoch": 2.70918506405015,
      "grad_norm": 0.13665929436683655,
      "learning_rate": 9.540956398368093e-05,
      "loss": 1.3846,
      "mean_token_accuracy": 0.7379218608140945,
      "num_tokens": 714380213.0,
      "step": 9940
    },
    {
      "entropy": 0.948689442873001,
      "epoch": 2.7119106023439628,
      "grad_norm": 0.09547930210828781,
      "learning_rate": 9.539698066203036e-05,
      "loss": 1.3867,
      "mean_token_accuracy": 0.7462014019489288,
      "num_tokens": 715100473.0,
      "step": 9950
    },
    {
      "entropy": 0.9547126933932304,
      "epoch": 2.7146361406377757,
      "grad_norm": 0.4112579822540283,
      "learning_rate": 9.538438094937028e-05,
      "loss": 1.3794,
      "mean_token_accuracy": 0.7413916006684304,
      "num_tokens": 715814075.0,
      "step": 9960
    },
    {
      "entropy": 0.9566810131072998,
      "epoch": 2.717361678931589,
      "grad_norm": 0.12511007487773895,
      "learning_rate": 9.537176485024993e-05,
      "loss": 1.3916,
      "mean_token_accuracy": 0.7371821865439415,
      "num_tokens": 716518360.0,
      "step": 9970
    },
    {
      "entropy": 0.9447938576340675,
      "epoch": 2.720087217225402,
      "grad_norm": 0.09474101662635803,
      "learning_rate": 9.535913236922447e-05,
      "loss": 1.3746,
      "mean_token_accuracy": 0.7438409894704818,
      "num_tokens": 717242082.0,
      "step": 9980
    },
    {
      "entropy": 0.9429836019873619,
      "epoch": 2.722812755519215,
      "grad_norm": 0.10164990276098251,
      "learning_rate": 9.534648351085498e-05,
      "loss": 1.3712,
      "mean_token_accuracy": 0.748205628991127,
      "num_tokens": 717970750.0,
      "step": 9990
    },
    {
      "entropy": 0.9465658456087113,
      "epoch": 2.725538293813028,
      "grad_norm": 0.09238191694021225,
      "learning_rate": 9.533381827970844e-05,
      "loss": 1.3776,
      "mean_token_accuracy": 0.7437403619289398,
      "num_tokens": 718692172.0,
      "step": 10000
    },
    {
      "entropy": 0.9468672856688499,
      "epoch": 2.7282638321068413,
      "grad_norm": 0.09535587579011917,
      "learning_rate": 9.532113668035776e-05,
      "loss": 1.3827,
      "mean_token_accuracy": 0.7401351481676102,
      "num_tokens": 719398745.0,
      "step": 10010
    },
    {
      "entropy": 0.9498398303985596,
      "epoch": 2.7309893704006543,
      "grad_norm": 0.09753476083278656,
      "learning_rate": 9.530843871738175e-05,
      "loss": 1.375,
      "mean_token_accuracy": 0.7446337208151818,
      "num_tokens": 720130026.0,
      "step": 10020
    },
    {
      "entropy": 0.9583035558462143,
      "epoch": 2.733714908694467,
      "grad_norm": 0.10755791515111923,
      "learning_rate": 9.529572439536513e-05,
      "loss": 1.3865,
      "mean_token_accuracy": 0.7396931052207947,
      "num_tokens": 720845902.0,
      "step": 10030
    },
    {
      "entropy": 0.9478771984577179,
      "epoch": 2.73644044698828,
      "grad_norm": 0.10072796791791916,
      "learning_rate": 9.528299371889855e-05,
      "loss": 1.3793,
      "mean_token_accuracy": 0.7432537838816643,
      "num_tokens": 721553882.0,
      "step": 10040
    },
    {
      "entropy": 0.9405006676912308,
      "epoch": 2.739165985282093,
      "grad_norm": 0.10096089541912079,
      "learning_rate": 9.527024669257849e-05,
      "loss": 1.3742,
      "mean_token_accuracy": 0.7456749126315116,
      "num_tokens": 722267845.0,
      "step": 10050
    },
    {
      "entropy": 0.9401244565844535,
      "epoch": 2.741891523575906,
      "grad_norm": 0.12743054330348969,
      "learning_rate": 9.525748332100743e-05,
      "loss": 1.3716,
      "mean_token_accuracy": 0.747305279970169,
      "num_tokens": 722988514.0,
      "step": 10060
    },
    {
      "entropy": 0.9473287984728813,
      "epoch": 2.7446170618697194,
      "grad_norm": 0.09784083813428879,
      "learning_rate": 9.524470360879368e-05,
      "loss": 1.385,
      "mean_token_accuracy": 0.744572938978672,
      "num_tokens": 723702201.0,
      "step": 10070
    },
    {
      "entropy": 0.949025246500969,
      "epoch": 2.7473426001635324,
      "grad_norm": 0.10419026017189026,
      "learning_rate": 9.523190756055147e-05,
      "loss": 1.3806,
      "mean_token_accuracy": 0.7432950124144554,
      "num_tokens": 724418507.0,
      "step": 10080
    },
    {
      "entropy": 0.9548242285847663,
      "epoch": 2.7500681384573453,
      "grad_norm": 0.1104128360748291,
      "learning_rate": 9.521909518090096e-05,
      "loss": 1.3944,
      "mean_token_accuracy": 0.7405140653252602,
      "num_tokens": 725119968.0,
      "step": 10090
    },
    {
      "entropy": 0.9464596956968307,
      "epoch": 2.7527936767511583,
      "grad_norm": 0.11928076297044754,
      "learning_rate": 9.520626647446818e-05,
      "loss": 1.3753,
      "mean_token_accuracy": 0.7412620186805725,
      "num_tokens": 725828274.0,
      "step": 10100
    },
    {
      "entropy": 0.9500831842422486,
      "epoch": 2.7555192150449717,
      "grad_norm": 0.08352699130773544,
      "learning_rate": 9.519342144588503e-05,
      "loss": 1.384,
      "mean_token_accuracy": 0.7443070754408836,
      "num_tokens": 726551641.0,
      "step": 10110
    },
    {
      "entropy": 0.9436436563730239,
      "epoch": 2.7582447533387846,
      "grad_norm": 0.0821971595287323,
      "learning_rate": 9.518056009978935e-05,
      "loss": 1.3717,
      "mean_token_accuracy": 0.7473061949014663,
      "num_tokens": 727282712.0,
      "step": 10120
    },
    {
      "entropy": 0.947759947180748,
      "epoch": 2.7609702916325976,
      "grad_norm": 0.09341433644294739,
      "learning_rate": 9.516768244082486e-05,
      "loss": 1.3953,
      "mean_token_accuracy": 0.7413911446928978,
      "num_tokens": 727987467.0,
      "step": 10130
    },
    {
      "entropy": 0.951104485988617,
      "epoch": 2.7636958299264105,
      "grad_norm": 0.08259022235870361,
      "learning_rate": 9.515478847364112e-05,
      "loss": 1.3734,
      "mean_token_accuracy": 0.7434945836663246,
      "num_tokens": 728712517.0,
      "step": 10140
    },
    {
      "entropy": 0.9454701945185662,
      "epoch": 2.7664213682202234,
      "grad_norm": 0.09898059070110321,
      "learning_rate": 9.514187820289367e-05,
      "loss": 1.379,
      "mean_token_accuracy": 0.7465288504958153,
      "num_tokens": 729438650.0,
      "step": 10150
    },
    {
      "entropy": 0.9487116679549217,
      "epoch": 2.7691469065140364,
      "grad_norm": 0.12520450353622437,
      "learning_rate": 9.512895163324383e-05,
      "loss": 1.3811,
      "mean_token_accuracy": 0.7442427203059196,
      "num_tokens": 730157457.0,
      "step": 10160
    },
    {
      "entropy": 0.9486458107829094,
      "epoch": 2.7718724448078493,
      "grad_norm": 0.13923697173595428,
      "learning_rate": 9.511600876935891e-05,
      "loss": 1.3778,
      "mean_token_accuracy": 0.7460005581378937,
      "num_tokens": 730880143.0,
      "step": 10170
    },
    {
      "entropy": 0.9487900391221047,
      "epoch": 2.7745979831016627,
      "grad_norm": 0.10820474475622177,
      "learning_rate": 9.510304961591203e-05,
      "loss": 1.3767,
      "mean_token_accuracy": 0.7412540033459664,
      "num_tokens": 731596819.0,
      "step": 10180
    },
    {
      "entropy": 0.9560275942087173,
      "epoch": 2.7773235213954757,
      "grad_norm": 0.10454489290714264,
      "learning_rate": 9.509007417758222e-05,
      "loss": 1.389,
      "mean_token_accuracy": 0.7426157101988793,
      "num_tokens": 732324778.0,
      "step": 10190
    },
    {
      "entropy": 0.9486469194293022,
      "epoch": 2.7800490596892886,
      "grad_norm": 0.18831004202365875,
      "learning_rate": 9.50770824590544e-05,
      "loss": 1.3813,
      "mean_token_accuracy": 0.7447981432080268,
      "num_tokens": 733047659.0,
      "step": 10200
    },
    {
      "entropy": 0.9435642853379249,
      "epoch": 2.7827745979831016,
      "grad_norm": 0.14489497244358063,
      "learning_rate": 9.50640744650193e-05,
      "loss": 1.3731,
      "mean_token_accuracy": 0.7452139586210251,
      "num_tokens": 733775297.0,
      "step": 10210
    },
    {
      "entropy": 0.9443108022212983,
      "epoch": 2.785500136276915,
      "grad_norm": 0.17383357882499695,
      "learning_rate": 9.505105020017364e-05,
      "loss": 1.3758,
      "mean_token_accuracy": 0.7443449541926384,
      "num_tokens": 734498209.0,
      "step": 10220
    },
    {
      "entropy": 0.9533507496118545,
      "epoch": 2.788225674570728,
      "grad_norm": 0.12784239649772644,
      "learning_rate": 9.503800966921993e-05,
      "loss": 1.3855,
      "mean_token_accuracy": 0.7430297493934631,
      "num_tokens": 735211652.0,
      "step": 10230
    },
    {
      "entropy": 0.9407243207097054,
      "epoch": 2.790951212864541,
      "grad_norm": 0.24755477905273438,
      "learning_rate": 9.502495287686657e-05,
      "loss": 1.3791,
      "mean_token_accuracy": 0.7442964062094688,
      "num_tokens": 735920070.0,
      "step": 10240
    },
    {
      "entropy": 0.9450778156518936,
      "epoch": 2.793676751158354,
      "grad_norm": 0.13384370505809784,
      "learning_rate": 9.501187982782785e-05,
      "loss": 1.3728,
      "mean_token_accuracy": 0.741634002327919,
      "num_tokens": 736627988.0,
      "step": 10250
    },
    {
      "entropy": 0.943241237103939,
      "epoch": 2.7964022894521667,
      "grad_norm": 0.09825392812490463,
      "learning_rate": 9.499879052682392e-05,
      "loss": 1.3785,
      "mean_token_accuracy": 0.7429873079061509,
      "num_tokens": 737338990.0,
      "step": 10260
    },
    {
      "entropy": 0.9448518842458725,
      "epoch": 2.7991278277459797,
      "grad_norm": 0.09042879939079285,
      "learning_rate": 9.498568497858079e-05,
      "loss": 1.3715,
      "mean_token_accuracy": 0.7506170704960823,
      "num_tokens": 738080465.0,
      "step": 10270
    },
    {
      "entropy": 0.9608897134661675,
      "epoch": 2.8018533660397926,
      "grad_norm": 0.09758147597312927,
      "learning_rate": 9.497256318783033e-05,
      "loss": 1.3893,
      "mean_token_accuracy": 0.7390094771981239,
      "num_tokens": 738792587.0,
      "step": 10280
    },
    {
      "entropy": 0.9479752659797669,
      "epoch": 2.804578904333606,
      "grad_norm": 0.12299107015132904,
      "learning_rate": 9.495942515931032e-05,
      "loss": 1.3758,
      "mean_token_accuracy": 0.7490738824009895,
      "num_tokens": 739524633.0,
      "step": 10290
    },
    {
      "entropy": 0.9539648801088333,
      "epoch": 2.807304442627419,
      "grad_norm": 0.09575517475605011,
      "learning_rate": 9.494627089776431e-05,
      "loss": 1.3843,
      "mean_token_accuracy": 0.7454695031046867,
      "num_tokens": 740252573.0,
      "step": 10300
    },
    {
      "entropy": 0.9500061124563217,
      "epoch": 2.810029980921232,
      "grad_norm": 0.11410588026046753,
      "learning_rate": 9.493310040794183e-05,
      "loss": 1.3912,
      "mean_token_accuracy": 0.7406770437955856,
      "num_tokens": 740963929.0,
      "step": 10310
    },
    {
      "entropy": 0.9521198511123657,
      "epoch": 2.812755519215045,
      "grad_norm": 0.09219717234373093,
      "learning_rate": 9.49199136945982e-05,
      "loss": 1.3793,
      "mean_token_accuracy": 0.7472849547863006,
      "num_tokens": 741697463.0,
      "step": 10320
    },
    {
      "entropy": 0.9477602332830429,
      "epoch": 2.8154810575088582,
      "grad_norm": 0.09398355334997177,
      "learning_rate": 9.490671076249458e-05,
      "loss": 1.3856,
      "mean_token_accuracy": 0.7394422218203545,
      "num_tokens": 742393454.0,
      "step": 10330
    },
    {
      "entropy": 0.9538507401943207,
      "epoch": 2.818206595802671,
      "grad_norm": 0.09424974024295807,
      "learning_rate": 9.489349161639805e-05,
      "loss": 1.3872,
      "mean_token_accuracy": 0.7407204285264015,
      "num_tokens": 743096713.0,
      "step": 10340
    },
    {
      "entropy": 0.9477018043398857,
      "epoch": 2.820932134096484,
      "grad_norm": 0.09561866521835327,
      "learning_rate": 9.488025626108147e-05,
      "loss": 1.379,
      "mean_token_accuracy": 0.74570182710886,
      "num_tokens": 743826659.0,
      "step": 10350
    },
    {
      "entropy": 0.9481017872691154,
      "epoch": 2.823657672390297,
      "grad_norm": 0.0974368155002594,
      "learning_rate": 9.486700470132361e-05,
      "loss": 1.3836,
      "mean_token_accuracy": 0.739516082406044,
      "num_tokens": 744532914.0,
      "step": 10360
    },
    {
      "entropy": 0.9480232343077659,
      "epoch": 2.82638321068411,
      "grad_norm": 0.12688322365283966,
      "learning_rate": 9.485373694190909e-05,
      "loss": 1.3805,
      "mean_token_accuracy": 0.7399832800030708,
      "num_tokens": 745248229.0,
      "step": 10370
    },
    {
      "entropy": 0.9455208957195282,
      "epoch": 2.829108748977923,
      "grad_norm": 0.1293535977602005,
      "learning_rate": 9.484045298762832e-05,
      "loss": 1.3767,
      "mean_token_accuracy": 0.744144344329834,
      "num_tokens": 745971792.0,
      "step": 10380
    },
    {
      "entropy": 0.9517705485224723,
      "epoch": 2.831834287271736,
      "grad_norm": 0.18343980610370636,
      "learning_rate": 9.482715284327762e-05,
      "loss": 1.3861,
      "mean_token_accuracy": 0.7427744567394257,
      "num_tokens": 746698383.0,
      "step": 10390
    },
    {
      "entropy": 0.9489703759551048,
      "epoch": 2.8345598255655493,
      "grad_norm": 0.09908950328826904,
      "learning_rate": 9.481383651365912e-05,
      "loss": 1.3823,
      "mean_token_accuracy": 0.7413738742470741,
      "num_tokens": 747420536.0,
      "step": 10400
    },
    {
      "entropy": 0.9515311121940613,
      "epoch": 2.8372853638593623,
      "grad_norm": 0.10469796508550644,
      "learning_rate": 9.480050400358083e-05,
      "loss": 1.3776,
      "mean_token_accuracy": 0.7367751017212868,
      "num_tokens": 748120021.0,
      "step": 10410
    },
    {
      "entropy": 0.9512437969446182,
      "epoch": 2.840010902153175,
      "grad_norm": 0.11225511133670807,
      "learning_rate": 9.478715531785657e-05,
      "loss": 1.3863,
      "mean_token_accuracy": 0.7434330567717552,
      "num_tokens": 748836539.0,
      "step": 10420
    },
    {
      "entropy": 0.9515815705060959,
      "epoch": 2.842736440446988,
      "grad_norm": 0.09097540378570557,
      "learning_rate": 9.477379046130598e-05,
      "loss": 1.3875,
      "mean_token_accuracy": 0.7462526351213455,
      "num_tokens": 749561315.0,
      "step": 10430
    },
    {
      "entropy": 0.9418442562222481,
      "epoch": 2.8454619787408015,
      "grad_norm": 0.10939962416887283,
      "learning_rate": 9.476040943875462e-05,
      "loss": 1.3744,
      "mean_token_accuracy": 0.7481460407376289,
      "num_tokens": 750278805.0,
      "step": 10440
    },
    {
      "entropy": 0.9443031251430511,
      "epoch": 2.8481875170346145,
      "grad_norm": 0.09361743181943893,
      "learning_rate": 9.474701225503381e-05,
      "loss": 1.3746,
      "mean_token_accuracy": 0.7502325609326362,
      "num_tokens": 751024327.0,
      "step": 10450
    },
    {
      "entropy": 0.9572816222906113,
      "epoch": 2.8509130553284274,
      "grad_norm": 0.0918792262673378,
      "learning_rate": 9.473359891498071e-05,
      "loss": 1.3801,
      "mean_token_accuracy": 0.7418233335018158,
      "num_tokens": 751749132.0,
      "step": 10460
    },
    {
      "entropy": 0.9410624951124191,
      "epoch": 2.8536385936222404,
      "grad_norm": 0.09585850685834885,
      "learning_rate": 9.472016942343836e-05,
      "loss": 1.3802,
      "mean_token_accuracy": 0.7431176856160164,
      "num_tokens": 752454813.0,
      "step": 10470
    },
    {
      "entropy": 0.9494479656219482,
      "epoch": 2.8563641319160533,
      "grad_norm": 0.10246778279542923,
      "learning_rate": 9.470672378525559e-05,
      "loss": 1.3788,
      "mean_token_accuracy": 0.7460354894399643,
      "num_tokens": 753182182.0,
      "step": 10480
    },
    {
      "entropy": 0.9505635082721711,
      "epoch": 2.8590896702098663,
      "grad_norm": 0.1411740630865097,
      "learning_rate": 9.469326200528709e-05,
      "loss": 1.3752,
      "mean_token_accuracy": 0.7432245314121246,
      "num_tokens": 753909834.0,
      "step": 10490
    },
    {
      "entropy": 0.9416063576936722,
      "epoch": 2.861815208503679,
      "grad_norm": 0.11739660799503326,
      "learning_rate": 9.467978408839335e-05,
      "loss": 1.3799,
      "mean_token_accuracy": 0.743518528342247,
      "num_tokens": 754614566.0,
      "step": 10500
    },
    {
      "entropy": 0.9634650573134422,
      "epoch": 2.8645407467974926,
      "grad_norm": 0.15798868238925934,
      "learning_rate": 9.46662900394407e-05,
      "loss": 1.3984,
      "mean_token_accuracy": 0.7380047783255577,
      "num_tokens": 755318065.0,
      "step": 10510
    },
    {
      "entropy": 0.943720032274723,
      "epoch": 2.8672662850913055,
      "grad_norm": 0.12477879971265793,
      "learning_rate": 9.46527798633013e-05,
      "loss": 1.3745,
      "mean_token_accuracy": 0.7484955221414566,
      "num_tokens": 756047541.0,
      "step": 10520
    },
    {
      "entropy": 0.9561996191740036,
      "epoch": 2.8699918233851185,
      "grad_norm": 0.10567452013492584,
      "learning_rate": 9.463925356485313e-05,
      "loss": 1.3872,
      "mean_token_accuracy": 0.7468387603759765,
      "num_tokens": 756779859.0,
      "step": 10530
    },
    {
      "entropy": 0.9433674648404121,
      "epoch": 2.8727173616789314,
      "grad_norm": 0.12189490348100662,
      "learning_rate": 9.462571114897998e-05,
      "loss": 1.3747,
      "mean_token_accuracy": 0.7466014489531517,
      "num_tokens": 757506757.0,
      "step": 10540
    },
    {
      "entropy": 0.9495260760188102,
      "epoch": 2.875442899972745,
      "grad_norm": 0.10370657593011856,
      "learning_rate": 9.461215262057146e-05,
      "loss": 1.382,
      "mean_token_accuracy": 0.7441823378205299,
      "num_tokens": 758236473.0,
      "step": 10550
    },
    {
      "entropy": 0.9483583778142929,
      "epoch": 2.8781684382665578,
      "grad_norm": 0.12287358194589615,
      "learning_rate": 9.459857798452303e-05,
      "loss": 1.3771,
      "mean_token_accuracy": 0.7479981765151024,
      "num_tokens": 758966654.0,
      "step": 10560
    },
    {
      "entropy": 0.9491520434617996,
      "epoch": 2.8808939765603707,
      "grad_norm": 0.08995507657527924,
      "learning_rate": 9.458498724573591e-05,
      "loss": 1.3908,
      "mean_token_accuracy": 0.7430084392428398,
      "num_tokens": 759678929.0,
      "step": 10570
    },
    {
      "entropy": 0.9527735188603401,
      "epoch": 2.8836195148541837,
      "grad_norm": 0.1153775081038475,
      "learning_rate": 9.45713804091172e-05,
      "loss": 1.3872,
      "mean_token_accuracy": 0.7407803103327751,
      "num_tokens": 760381545.0,
      "step": 10580
    },
    {
      "entropy": 0.9517191231250763,
      "epoch": 2.8863450531479966,
      "grad_norm": 0.10239632427692413,
      "learning_rate": 9.455775747957973e-05,
      "loss": 1.3807,
      "mean_token_accuracy": 0.7425818264484405,
      "num_tokens": 761101701.0,
      "step": 10590
    },
    {
      "entropy": 0.9448929950594902,
      "epoch": 2.8890705914418096,
      "grad_norm": 0.2964629530906677,
      "learning_rate": 9.454411846204224e-05,
      "loss": 1.3881,
      "mean_token_accuracy": 0.7405307441949844,
      "num_tokens": 761802896.0,
      "step": 10600
    },
    {
      "entropy": 0.9530297860503196,
      "epoch": 2.891796129735623,
      "grad_norm": 0.14000527560710907,
      "learning_rate": 9.45304633614292e-05,
      "loss": 1.3876,
      "mean_token_accuracy": 0.7408836930990219,
      "num_tokens": 762524473.0,
      "step": 10610
    },
    {
      "entropy": 0.9478141114115715,
      "epoch": 2.894521668029436,
      "grad_norm": 0.11951442807912827,
      "learning_rate": 9.451679218267092e-05,
      "loss": 1.3801,
      "mean_token_accuracy": 0.7429118230938911,
      "num_tokens": 763232538.0,
      "step": 10620
    },
    {
      "entropy": 0.9559646502137185,
      "epoch": 2.897247206323249,
      "grad_norm": 0.09150029718875885,
      "learning_rate": 9.450310493070349e-05,
      "loss": 1.3882,
      "mean_token_accuracy": 0.7428459390997887,
      "num_tokens": 763954109.0,
      "step": 10630
    },
    {
      "entropy": 0.9576432898640632,
      "epoch": 2.899972744617062,
      "grad_norm": 0.11356475949287415,
      "learning_rate": 9.448940161046884e-05,
      "loss": 1.3928,
      "mean_token_accuracy": 0.7371211186051368,
      "num_tokens": 764670941.0,
      "step": 10640
    },
    {
      "entropy": 0.9471401020884513,
      "epoch": 2.902698282910875,
      "grad_norm": 0.09661433845758438,
      "learning_rate": 9.447568222691467e-05,
      "loss": 1.3802,
      "mean_token_accuracy": 0.7479994982481003,
      "num_tokens": 765398023.0,
      "step": 10650
    },
    {
      "entropy": 0.9421188428997993,
      "epoch": 2.905423821204688,
      "grad_norm": 0.09525974839925766,
      "learning_rate": 9.446194678499451e-05,
      "loss": 1.3724,
      "mean_token_accuracy": 0.7448062315583229,
      "num_tokens": 766103654.0,
      "step": 10660
    },
    {
      "entropy": 0.9504467383027076,
      "epoch": 2.908149359498501,
      "grad_norm": 0.07963963598012924,
      "learning_rate": 9.444819528966766e-05,
      "loss": 1.3795,
      "mean_token_accuracy": 0.7468686282634736,
      "num_tokens": 766838816.0,
      "step": 10670
    },
    {
      "entropy": 0.9504798486828804,
      "epoch": 2.910874897792314,
      "grad_norm": 0.12807469069957733,
      "learning_rate": 9.443442774589925e-05,
      "loss": 1.3799,
      "mean_token_accuracy": 0.744924345612526,
      "num_tokens": 767564829.0,
      "step": 10680
    },
    {
      "entropy": 0.9473272353410721,
      "epoch": 2.913600436086127,
      "grad_norm": 0.08859226107597351,
      "learning_rate": 9.442064415866015e-05,
      "loss": 1.3785,
      "mean_token_accuracy": 0.7422213032841682,
      "num_tokens": 768278005.0,
      "step": 10690
    },
    {
      "entropy": 0.9543716087937355,
      "epoch": 2.91632597437994,
      "grad_norm": 0.11691927909851074,
      "learning_rate": 9.440684453292706e-05,
      "loss": 1.3888,
      "mean_token_accuracy": 0.7412212237715721,
      "num_tokens": 768987600.0,
      "step": 10700
    },
    {
      "entropy": 0.9488771393895149,
      "epoch": 2.919051512673753,
      "grad_norm": 0.08694035559892654,
      "learning_rate": 9.439302887368249e-05,
      "loss": 1.3779,
      "mean_token_accuracy": 0.7432838320732117,
      "num_tokens": 769715397.0,
      "step": 10710
    },
    {
      "entropy": 0.9454629391431808,
      "epoch": 2.9217770509675662,
      "grad_norm": 0.10040954500436783,
      "learning_rate": 9.437919718591468e-05,
      "loss": 1.3793,
      "mean_token_accuracy": 0.7399373590946198,
      "num_tokens": 770429390.0,
      "step": 10720
    },
    {
      "entropy": 0.9490958735346794,
      "epoch": 2.924502589261379,
      "grad_norm": 0.10891756415367126,
      "learning_rate": 9.436534947461772e-05,
      "loss": 1.3774,
      "mean_token_accuracy": 0.7455389559268951,
      "num_tokens": 771163171.0,
      "step": 10730
    },
    {
      "entropy": 0.9498210698366165,
      "epoch": 2.927228127555192,
      "grad_norm": 0.09342612326145172,
      "learning_rate": 9.435148574479144e-05,
      "loss": 1.3779,
      "mean_token_accuracy": 0.738006317615509,
      "num_tokens": 771860027.0,
      "step": 10740
    },
    {
      "entropy": 0.9461888819932938,
      "epoch": 2.929953665849005,
      "grad_norm": 0.09771758317947388,
      "learning_rate": 9.43376060014415e-05,
      "loss": 1.3776,
      "mean_token_accuracy": 0.7404180765151978,
      "num_tokens": 772567486.0,
      "step": 10750
    },
    {
      "entropy": 0.9388180360198021,
      "epoch": 2.9326792041428185,
      "grad_norm": 0.10150643438100815,
      "learning_rate": 9.432371024957927e-05,
      "loss": 1.3753,
      "mean_token_accuracy": 0.7463704735040665,
      "num_tokens": 773289854.0,
      "step": 10760
    },
    {
      "entropy": 0.9550876408815384,
      "epoch": 2.9354047424366314,
      "grad_norm": 0.09483397006988525,
      "learning_rate": 9.430979849422196e-05,
      "loss": 1.3819,
      "mean_token_accuracy": 0.7382730200886727,
      "num_tokens": 774003702.0,
      "step": 10770
    },
    {
      "entropy": 0.9492468416690827,
      "epoch": 2.9381302807304444,
      "grad_norm": 0.11253222823143005,
      "learning_rate": 9.429587074039254e-05,
      "loss": 1.3859,
      "mean_token_accuracy": 0.7406412959098816,
      "num_tokens": 774715012.0,
      "step": 10780
    },
    {
      "entropy": 0.9470512971282006,
      "epoch": 2.9408558190242573,
      "grad_norm": 0.1132352203130722,
      "learning_rate": 9.428192699311979e-05,
      "loss": 1.3757,
      "mean_token_accuracy": 0.7429844021797181,
      "num_tokens": 775431299.0,
      "step": 10790
    },
    {
      "entropy": 0.9407175153493881,
      "epoch": 2.9435813573180702,
      "grad_norm": 0.11837431788444519,
      "learning_rate": 9.426796725743818e-05,
      "loss": 1.3696,
      "mean_token_accuracy": 0.7453804090619087,
      "num_tokens": 776162306.0,
      "step": 10800
    },
    {
      "entropy": 0.9463397905230522,
      "epoch": 2.946306895611883,
      "grad_norm": 0.14300303161144257,
      "learning_rate": 9.425399153838804e-05,
      "loss": 1.3802,
      "mean_token_accuracy": 0.7406564101576805,
      "num_tokens": 776875978.0,
      "step": 10810
    },
    {
      "entropy": 0.9498815953731536,
      "epoch": 2.949032433905696,
      "grad_norm": 0.1136244460940361,
      "learning_rate": 9.423999984101542e-05,
      "loss": 1.3762,
      "mean_token_accuracy": 0.7416051119565964,
      "num_tokens": 777590020.0,
      "step": 10820
    },
    {
      "entropy": 0.9428626239299774,
      "epoch": 2.9517579721995095,
      "grad_norm": 0.10531148314476013,
      "learning_rate": 9.422599217037213e-05,
      "loss": 1.3705,
      "mean_token_accuracy": 0.74758260846138,
      "num_tokens": 778321873.0,
      "step": 10830
    },
    {
      "entropy": 0.9489976704120636,
      "epoch": 2.9544835104933225,
      "grad_norm": 0.13479472696781158,
      "learning_rate": 9.421196853151584e-05,
      "loss": 1.38,
      "mean_token_accuracy": 0.7459505274891853,
      "num_tokens": 779049676.0,
      "step": 10840
    },
    {
      "entropy": 0.9520537599921226,
      "epoch": 2.9572090487871354,
      "grad_norm": 0.08216880261898041,
      "learning_rate": 9.419792892950987e-05,
      "loss": 1.386,
      "mean_token_accuracy": 0.7429448261857032,
      "num_tokens": 779775404.0,
      "step": 10850
    },
    {
      "entropy": 0.949109499156475,
      "epoch": 2.9599345870809484,
      "grad_norm": 0.1596001535654068,
      "learning_rate": 9.418387336942335e-05,
      "loss": 1.3752,
      "mean_token_accuracy": 0.7430404499173164,
      "num_tokens": 780497904.0,
      "step": 10860
    },
    {
      "entropy": 0.9505121961236,
      "epoch": 2.9626601253747618,
      "grad_norm": 0.11224663257598877,
      "learning_rate": 9.416980185633117e-05,
      "loss": 1.3775,
      "mean_token_accuracy": 0.7451267004013061,
      "num_tokens": 781217060.0,
      "step": 10870
    },
    {
      "entropy": 0.9447997346520424,
      "epoch": 2.9653856636685747,
      "grad_norm": 0.11577615141868591,
      "learning_rate": 9.415571439531402e-05,
      "loss": 1.3751,
      "mean_token_accuracy": 0.744077518582344,
      "num_tokens": 781931549.0,
      "step": 10880
    },
    {
      "entropy": 0.9436752960085869,
      "epoch": 2.9681112019623876,
      "grad_norm": 0.153329536318779,
      "learning_rate": 9.414161099145828e-05,
      "loss": 1.3806,
      "mean_token_accuracy": 0.7423443600535393,
      "num_tokens": 782623152.0,
      "step": 10890
    },
    {
      "entropy": 0.9503539055585861,
      "epoch": 2.9708367402562006,
      "grad_norm": 0.19156911969184875,
      "learning_rate": 9.412749164985611e-05,
      "loss": 1.383,
      "mean_token_accuracy": 0.7420576244592667,
      "num_tokens": 783346287.0,
      "step": 10900
    },
    {
      "entropy": 0.9467993035912514,
      "epoch": 2.9735622785500135,
      "grad_norm": 0.09959796071052551,
      "learning_rate": 9.411335637560546e-05,
      "loss": 1.3777,
      "mean_token_accuracy": 0.7447941780090332,
      "num_tokens": 784068242.0,
      "step": 10910
    },
    {
      "entropy": 0.9460756421089173,
      "epoch": 2.9762878168438265,
      "grad_norm": 0.09663764387369156,
      "learning_rate": 9.409920517380998e-05,
      "loss": 1.3829,
      "mean_token_accuracy": 0.743578876554966,
      "num_tokens": 784785982.0,
      "step": 10920
    },
    {
      "entropy": 0.9464081883430481,
      "epoch": 2.9790133551376394,
      "grad_norm": 0.098046213388443,
      "learning_rate": 9.408503804957912e-05,
      "loss": 1.3731,
      "mean_token_accuracy": 0.7468303069472313,
      "num_tokens": 785524509.0,
      "step": 10930
    },
    {
      "entropy": 0.9433768823742866,
      "epoch": 2.981738893431453,
      "grad_norm": 0.08793983608484268,
      "learning_rate": 9.407085500802804e-05,
      "loss": 1.3731,
      "mean_token_accuracy": 0.7433084145188331,
      "num_tokens": 786241993.0,
      "step": 10940
    },
    {
      "entropy": 0.9438702210783958,
      "epoch": 2.9844644317252658,
      "grad_norm": 0.11304867267608643,
      "learning_rate": 9.405665605427764e-05,
      "loss": 1.3808,
      "mean_token_accuracy": 0.7470975622534752,
      "num_tokens": 786966806.0,
      "step": 10950
    },
    {
      "entropy": 0.9451516225934029,
      "epoch": 2.9871899700190787,
      "grad_norm": 0.09482454508543015,
      "learning_rate": 9.404244119345462e-05,
      "loss": 1.3702,
      "mean_token_accuracy": 0.7435500398278236,
      "num_tokens": 787685395.0,
      "step": 10960
    },
    {
      "entropy": 0.9512408003211021,
      "epoch": 2.9899155083128917,
      "grad_norm": 0.10392162203788757,
      "learning_rate": 9.402821043069137e-05,
      "loss": 1.3828,
      "mean_token_accuracy": 0.7425945445895195,
      "num_tokens": 788402117.0,
      "step": 10970
    },
    {
      "entropy": 0.9488069012761116,
      "epoch": 2.992641046606705,
      "grad_norm": 0.08627504110336304,
      "learning_rate": 9.401396377112605e-05,
      "loss": 1.3775,
      "mean_token_accuracy": 0.7512379825115204,
      "num_tokens": 789144727.0,
      "step": 10980
    },
    {
      "entropy": 0.9501185655593872,
      "epoch": 2.995366584900518,
      "grad_norm": 0.10031311959028244,
      "learning_rate": 9.399970121990258e-05,
      "loss": 1.3763,
      "mean_token_accuracy": 0.7431940719485283,
      "num_tokens": 789864544.0,
      "step": 10990
    },
    {
      "entropy": 0.9512557774782181,
      "epoch": 2.998092123194331,
      "grad_norm": 0.10265135020017624,
      "learning_rate": 9.398542278217054e-05,
      "loss": 1.3793,
      "mean_token_accuracy": 0.7444485172629356,
      "num_tokens": 790586485.0,
      "step": 11000
    },
    {
      "entropy": 0.9507931903004646,
      "epoch": 3.000817661488144,
      "grad_norm": 0.0883914902806282,
      "learning_rate": 9.39711284630853e-05,
      "loss": 1.3836,
      "mean_token_accuracy": 0.745827704668045,
      "num_tokens": 791321126.0,
      "step": 11010
    },
    {
      "entropy": 0.9451069980859756,
      "epoch": 3.003543199781957,
      "grad_norm": 0.09480870515108109,
      "learning_rate": 9.395681826780798e-05,
      "loss": 1.3696,
      "mean_token_accuracy": 0.7437042579054832,
      "num_tokens": 792042747.0,
      "step": 11020
    },
    {
      "entropy": 0.9346920728683472,
      "epoch": 3.0062687380757698,
      "grad_norm": 0.11070021986961365,
      "learning_rate": 9.394249220150541e-05,
      "loss": 1.3606,
      "mean_token_accuracy": 0.750485347211361,
      "num_tokens": 792774577.0,
      "step": 11030
    },
    {
      "entropy": 0.935794934630394,
      "epoch": 3.008994276369583,
      "grad_norm": 0.08571377396583557,
      "learning_rate": 9.392815026935016e-05,
      "loss": 1.3649,
      "mean_token_accuracy": 0.7491550222039223,
      "num_tokens": 793512462.0,
      "step": 11040
    },
    {
      "entropy": 0.93894032984972,
      "epoch": 3.011719814663396,
      "grad_norm": 0.09834165871143341,
      "learning_rate": 9.39137924765205e-05,
      "loss": 1.3694,
      "mean_token_accuracy": 0.7463561862707138,
      "num_tokens": 794236781.0,
      "step": 11050
    },
    {
      "entropy": 0.9381979539990425,
      "epoch": 3.014445352957209,
      "grad_norm": 0.09578172862529755,
      "learning_rate": 9.389941882820045e-05,
      "loss": 1.3755,
      "mean_token_accuracy": 0.7448102459311485,
      "num_tokens": 794953031.0,
      "step": 11060
    },
    {
      "entropy": 0.9406249284744262,
      "epoch": 3.017170891251022,
      "grad_norm": 0.09311778843402863,
      "learning_rate": 9.388502932957975e-05,
      "loss": 1.3734,
      "mean_token_accuracy": 0.7449426963925362,
      "num_tokens": 795663779.0,
      "step": 11070
    },
    {
      "entropy": 0.9419813230633736,
      "epoch": 3.019896429544835,
      "grad_norm": 0.10084322839975357,
      "learning_rate": 9.38706239858539e-05,
      "loss": 1.3642,
      "mean_token_accuracy": 0.7455672308802604,
      "num_tokens": 796394650.0,
      "step": 11080
    },
    {
      "entropy": 0.9440831586718559,
      "epoch": 3.0226219678386483,
      "grad_norm": 0.10120902210474014,
      "learning_rate": 9.385620280222402e-05,
      "loss": 1.3681,
      "mean_token_accuracy": 0.7443292528390885,
      "num_tokens": 797127106.0,
      "step": 11090
    },
    {
      "entropy": 0.940793713927269,
      "epoch": 3.0253475061324613,
      "grad_norm": 0.1022334098815918,
      "learning_rate": 9.384176578389707e-05,
      "loss": 1.3745,
      "mean_token_accuracy": 0.74081010222435,
      "num_tokens": 797834893.0,
      "step": 11100
    },
    {
      "entropy": 0.9389016404747963,
      "epoch": 3.0280730444262742,
      "grad_norm": 0.08720928430557251,
      "learning_rate": 9.382731293608565e-05,
      "loss": 1.367,
      "mean_token_accuracy": 0.7487114503979683,
      "num_tokens": 798560747.0,
      "step": 11110
    },
    {
      "entropy": 0.9339886143803596,
      "epoch": 3.030798582720087,
      "grad_norm": 0.10605648905038834,
      "learning_rate": 9.38128442640081e-05,
      "loss": 1.3659,
      "mean_token_accuracy": 0.746621385216713,
      "num_tokens": 799276742.0,
      "step": 11120
    },
    {
      "entropy": 0.9361902475357056,
      "epoch": 3.0335241210139,
      "grad_norm": 0.0876423716545105,
      "learning_rate": 9.379835977288847e-05,
      "loss": 1.3665,
      "mean_token_accuracy": 0.7457955434918404,
      "num_tokens": 800002088.0,
      "step": 11130
    },
    {
      "entropy": 0.9462328448891639,
      "epoch": 3.036249659307713,
      "grad_norm": 0.09621265530586243,
      "learning_rate": 9.378385946795654e-05,
      "loss": 1.3718,
      "mean_token_accuracy": 0.7416390955448151,
      "num_tokens": 800711281.0,
      "step": 11140
    },
    {
      "entropy": 0.9392537325620651,
      "epoch": 3.0389751976015265,
      "grad_norm": 0.09904195368289948,
      "learning_rate": 9.376934335444774e-05,
      "loss": 1.3675,
      "mean_token_accuracy": 0.7491195678710938,
      "num_tokens": 801438797.0,
      "step": 11150
    },
    {
      "entropy": 0.9405299961566925,
      "epoch": 3.0417007358953394,
      "grad_norm": 0.08887912333011627,
      "learning_rate": 9.375481143760329e-05,
      "loss": 1.3805,
      "mean_token_accuracy": 0.7460344031453132,
      "num_tokens": 802155484.0,
      "step": 11160
    },
    {
      "entropy": 0.9410595312714577,
      "epoch": 3.0444262741891523,
      "grad_norm": 0.11672388762235641,
      "learning_rate": 9.374026372267008e-05,
      "loss": 1.379,
      "mean_token_accuracy": 0.7432695344090462,
      "num_tokens": 802862801.0,
      "step": 11170
    },
    {
      "entropy": 0.9330152377486229,
      "epoch": 3.0471518124829653,
      "grad_norm": 0.10320542007684708,
      "learning_rate": 9.372570021490067e-05,
      "loss": 1.3603,
      "mean_token_accuracy": 0.750064055621624,
      "num_tokens": 803584098.0,
      "step": 11180
    },
    {
      "entropy": 0.9365206196904182,
      "epoch": 3.0498773507767782,
      "grad_norm": 0.10336554795503616,
      "learning_rate": 9.371112091955337e-05,
      "loss": 1.3605,
      "mean_token_accuracy": 0.746023739874363,
      "num_tokens": 804309097.0,
      "step": 11190
    },
    {
      "entropy": 0.9407736912369729,
      "epoch": 3.0526028890705916,
      "grad_norm": 0.09537152200937271,
      "learning_rate": 9.369652584189216e-05,
      "loss": 1.3754,
      "mean_token_accuracy": 0.7431817069649697,
      "num_tokens": 805020667.0,
      "step": 11200
    },
    {
      "entropy": 0.9329175144433975,
      "epoch": 3.0553284273644046,
      "grad_norm": 0.11751585453748703,
      "learning_rate": 9.368191498718675e-05,
      "loss": 1.3714,
      "mean_token_accuracy": 0.7441901594400406,
      "num_tokens": 805725680.0,
      "step": 11210
    },
    {
      "entropy": 0.942389452457428,
      "epoch": 3.0580539656582175,
      "grad_norm": 0.09297958761453629,
      "learning_rate": 9.36672883607125e-05,
      "loss": 1.3726,
      "mean_token_accuracy": 0.7401214390993118,
      "num_tokens": 806434057.0,
      "step": 11220
    },
    {
      "entropy": 0.9488732665777206,
      "epoch": 3.0607795039520305,
      "grad_norm": 0.10751072317361832,
      "learning_rate": 9.365264596775051e-05,
      "loss": 1.377,
      "mean_token_accuracy": 0.7423231542110443,
      "num_tokens": 807150329.0,
      "step": 11230
    },
    {
      "entropy": 0.9424430325627327,
      "epoch": 3.0635050422458434,
      "grad_norm": 0.09440799802541733,
      "learning_rate": 9.363798781358757e-05,
      "loss": 1.3729,
      "mean_token_accuracy": 0.7408807426691055,
      "num_tokens": 807863155.0,
      "step": 11240
    },
    {
      "entropy": 0.9404341802001,
      "epoch": 3.0662305805396564,
      "grad_norm": 0.11174992471933365,
      "learning_rate": 9.362331390351613e-05,
      "loss": 1.379,
      "mean_token_accuracy": 0.7399561330676079,
      "num_tokens": 808568191.0,
      "step": 11250
    },
    {
      "entropy": 0.9390634030103684,
      "epoch": 3.0689561188334697,
      "grad_norm": 0.10013566166162491,
      "learning_rate": 9.360862424283433e-05,
      "loss": 1.3762,
      "mean_token_accuracy": 0.7432549253106118,
      "num_tokens": 809273011.0,
      "step": 11260
    },
    {
      "entropy": 0.9395414397120476,
      "epoch": 3.0716816571272827,
      "grad_norm": 0.21124373376369476,
      "learning_rate": 9.359391883684603e-05,
      "loss": 1.3657,
      "mean_token_accuracy": 0.7475090369582176,
      "num_tokens": 809997397.0,
      "step": 11270
    },
    {
      "entropy": 0.9396891310811043,
      "epoch": 3.0744071954210956,
      "grad_norm": 0.09619201719760895,
      "learning_rate": 9.357919769086074e-05,
      "loss": 1.3731,
      "mean_token_accuracy": 0.743741637468338,
      "num_tokens": 810701656.0,
      "step": 11280
    },
    {
      "entropy": 0.9452935412526131,
      "epoch": 3.0771327337149086,
      "grad_norm": 0.10465989261865616,
      "learning_rate": 9.356446081019369e-05,
      "loss": 1.3777,
      "mean_token_accuracy": 0.7458249852061272,
      "num_tokens": 811425807.0,
      "step": 11290
    },
    {
      "entropy": 0.9400225415825844,
      "epoch": 3.0798582720087215,
      "grad_norm": 0.1047290712594986,
      "learning_rate": 9.354970820016576e-05,
      "loss": 1.3673,
      "mean_token_accuracy": 0.748205590248108,
      "num_tokens": 812148517.0,
      "step": 11300
    },
    {
      "entropy": 0.9372500717639923,
      "epoch": 3.082583810302535,
      "grad_norm": 0.09082084149122238,
      "learning_rate": 9.353493986610351e-05,
      "loss": 1.3675,
      "mean_token_accuracy": 0.7476839691400528,
      "num_tokens": 812886158.0,
      "step": 11310
    },
    {
      "entropy": 0.9453914582729339,
      "epoch": 3.085309348596348,
      "grad_norm": 0.15638625621795654,
      "learning_rate": 9.35201558133392e-05,
      "loss": 1.3834,
      "mean_token_accuracy": 0.7417276456952095,
      "num_tokens": 813598294.0,
      "step": 11320
    },
    {
      "entropy": 0.9417727872729301,
      "epoch": 3.088034886890161,
      "grad_norm": 0.0936673954129219,
      "learning_rate": 9.350535604721074e-05,
      "loss": 1.3822,
      "mean_token_accuracy": 0.7389746978878975,
      "num_tokens": 814286851.0,
      "step": 11330
    },
    {
      "entropy": 0.9376664489507676,
      "epoch": 3.0907604251839738,
      "grad_norm": 0.09966736286878586,
      "learning_rate": 9.349054057306174e-05,
      "loss": 1.3723,
      "mean_token_accuracy": 0.742907702922821,
      "num_tokens": 815002766.0,
      "step": 11340
    },
    {
      "entropy": 0.9480553328990936,
      "epoch": 3.0934859634777867,
      "grad_norm": 0.11020088195800781,
      "learning_rate": 9.347570939624147e-05,
      "loss": 1.3777,
      "mean_token_accuracy": 0.7439572513103485,
      "num_tokens": 815720988.0,
      "step": 11350
    },
    {
      "entropy": 0.9363390222191811,
      "epoch": 3.0962115017716,
      "grad_norm": 0.11631142348051071,
      "learning_rate": 9.346086252210483e-05,
      "loss": 1.3608,
      "mean_token_accuracy": 0.7509266182780265,
      "num_tokens": 816459846.0,
      "step": 11360
    },
    {
      "entropy": 0.9384995147585868,
      "epoch": 3.098937040065413,
      "grad_norm": 0.10329711437225342,
      "learning_rate": 9.344599995601247e-05,
      "loss": 1.3635,
      "mean_token_accuracy": 0.7480807676911354,
      "num_tokens": 817198078.0,
      "step": 11370
    },
    {
      "entropy": 0.9449541583657265,
      "epoch": 3.101662578359226,
      "grad_norm": 0.12069828808307648,
      "learning_rate": 9.343112170333063e-05,
      "loss": 1.3711,
      "mean_token_accuracy": 0.7442483931779862,
      "num_tokens": 817920801.0,
      "step": 11380
    },
    {
      "entropy": 0.9456265062093735,
      "epoch": 3.104388116653039,
      "grad_norm": 0.09198126941919327,
      "learning_rate": 9.341622776943126e-05,
      "loss": 1.3784,
      "mean_token_accuracy": 0.7414027154445648,
      "num_tokens": 818628643.0,
      "step": 11390
    },
    {
      "entropy": 0.9380450159311294,
      "epoch": 3.107113654946852,
      "grad_norm": 0.11404696106910706,
      "learning_rate": 9.340131815969196e-05,
      "loss": 1.3688,
      "mean_token_accuracy": 0.741287750005722,
      "num_tokens": 819339287.0,
      "step": 11400
    },
    {
      "entropy": 0.9412304505705833,
      "epoch": 3.109839193240665,
      "grad_norm": 0.09746160358190536,
      "learning_rate": 9.338639287949598e-05,
      "loss": 1.37,
      "mean_token_accuracy": 0.7421133562922477,
      "num_tokens": 820052416.0,
      "step": 11410
    },
    {
      "entropy": 0.9427403524518013,
      "epoch": 3.112564731534478,
      "grad_norm": 0.13241012394428253,
      "learning_rate": 9.337145193423222e-05,
      "loss": 1.375,
      "mean_token_accuracy": 0.7437486976385117,
      "num_tokens": 820769288.0,
      "step": 11420
    },
    {
      "entropy": 0.9317566394805908,
      "epoch": 3.115290269828291,
      "grad_norm": 0.09343224763870239,
      "learning_rate": 9.335649532929526e-05,
      "loss": 1.3608,
      "mean_token_accuracy": 0.7504123464226723,
      "num_tokens": 821506352.0,
      "step": 11430
    },
    {
      "entropy": 0.939399915933609,
      "epoch": 3.118015808122104,
      "grad_norm": 0.09214864671230316,
      "learning_rate": 9.334152307008532e-05,
      "loss": 1.3734,
      "mean_token_accuracy": 0.7457444086670876,
      "num_tokens": 822229912.0,
      "step": 11440
    },
    {
      "entropy": 0.9384179800748825,
      "epoch": 3.120741346415917,
      "grad_norm": 0.10016916692256927,
      "learning_rate": 9.332653516200829e-05,
      "loss": 1.3721,
      "mean_token_accuracy": 0.7450738236308098,
      "num_tokens": 822944677.0,
      "step": 11450
    },
    {
      "entropy": 0.9418606579303741,
      "epoch": 3.12346688470973,
      "grad_norm": 0.09754331409931183,
      "learning_rate": 9.331153161047568e-05,
      "loss": 1.3664,
      "mean_token_accuracy": 0.7417466431856156,
      "num_tokens": 823654759.0,
      "step": 11460
    },
    {
      "entropy": 0.9352231919765472,
      "epoch": 3.1261924230035434,
      "grad_norm": 0.10757584124803543,
      "learning_rate": 9.329651242090468e-05,
      "loss": 1.3668,
      "mean_token_accuracy": 0.7431150436401367,
      "num_tokens": 824374173.0,
      "step": 11470
    },
    {
      "entropy": 0.9361462995409966,
      "epoch": 3.1289179612973563,
      "grad_norm": 0.1653221696615219,
      "learning_rate": 9.328147759871809e-05,
      "loss": 1.3722,
      "mean_token_accuracy": 0.7468035846948624,
      "num_tokens": 825083575.0,
      "step": 11480
    },
    {
      "entropy": 0.9438735976815223,
      "epoch": 3.1316434995911693,
      "grad_norm": 0.08508539944887161,
      "learning_rate": 9.326642714934439e-05,
      "loss": 1.3699,
      "mean_token_accuracy": 0.7426930829882622,
      "num_tokens": 825803337.0,
      "step": 11490
    },
    {
      "entropy": 0.9398536905646324,
      "epoch": 3.1343690378849822,
      "grad_norm": 0.12729324400424957,
      "learning_rate": 9.325136107821768e-05,
      "loss": 1.3646,
      "mean_token_accuracy": 0.7444761872291565,
      "num_tokens": 826521775.0,
      "step": 11500
    },
    {
      "entropy": 0.9386954724788665,
      "epoch": 3.137094576178795,
      "grad_norm": 0.13203135132789612,
      "learning_rate": 9.323627939077771e-05,
      "loss": 1.3716,
      "mean_token_accuracy": 0.7460088655352592,
      "num_tokens": 827251356.0,
      "step": 11510
    },
    {
      "entropy": 0.9365203708410264,
      "epoch": 3.1398201144726086,
      "grad_norm": 0.11855225265026093,
      "learning_rate": 9.322118209246987e-05,
      "loss": 1.3721,
      "mean_token_accuracy": 0.7442961215972901,
      "num_tokens": 827958502.0,
      "step": 11520
    },
    {
      "entropy": 0.9445111289620399,
      "epoch": 3.1425456527664215,
      "grad_norm": 0.10418067127466202,
      "learning_rate": 9.320606918874518e-05,
      "loss": 1.3717,
      "mean_token_accuracy": 0.7435276076197624,
      "num_tokens": 828686325.0,
      "step": 11530
    },
    {
      "entropy": 0.9411079570651054,
      "epoch": 3.1452711910602345,
      "grad_norm": 0.10798341035842896,
      "learning_rate": 9.319094068506029e-05,
      "loss": 1.3664,
      "mean_token_accuracy": 0.7485539555549622,
      "num_tokens": 829413364.0,
      "step": 11540
    },
    {
      "entropy": 0.9384330570697784,
      "epoch": 3.1479967293540474,
      "grad_norm": 0.09768079966306686,
      "learning_rate": 9.317579658687749e-05,
      "loss": 1.3708,
      "mean_token_accuracy": 0.747562299668789,
      "num_tokens": 830144904.0,
      "step": 11550
    },
    {
      "entropy": 0.9414130836725235,
      "epoch": 3.1507222676478603,
      "grad_norm": 0.10029740631580353,
      "learning_rate": 9.316063689966472e-05,
      "loss": 1.3683,
      "mean_token_accuracy": 0.7410704106092453,
      "num_tokens": 830857702.0,
      "step": 11560
    },
    {
      "entropy": 0.9360662370920181,
      "epoch": 3.1534478059416733,
      "grad_norm": 0.11149334162473679,
      "learning_rate": 9.314546162889551e-05,
      "loss": 1.3742,
      "mean_token_accuracy": 0.7429559230804443,
      "num_tokens": 831571151.0,
      "step": 11570
    },
    {
      "entropy": 0.9414043292403221,
      "epoch": 3.1561733442354867,
      "grad_norm": 0.1495295763015747,
      "learning_rate": 9.313027078004903e-05,
      "loss": 1.377,
      "mean_token_accuracy": 0.7445568561553955,
      "num_tokens": 832274229.0,
      "step": 11580
    },
    {
      "entropy": 0.9276083469390869,
      "epoch": 3.1588988825292996,
      "grad_norm": 0.09622129797935486,
      "learning_rate": 9.311506435861011e-05,
      "loss": 1.3577,
      "mean_token_accuracy": 0.747485676407814,
      "num_tokens": 832993745.0,
      "step": 11590
    },
    {
      "entropy": 0.9446370765566826,
      "epoch": 3.1616244208231126,
      "grad_norm": 0.09544003754854202,
      "learning_rate": 9.309984237006912e-05,
      "loss": 1.3857,
      "mean_token_accuracy": 0.7386672690510749,
      "num_tokens": 833686366.0,
      "step": 11600
    },
    {
      "entropy": 0.934392660856247,
      "epoch": 3.1643499591169255,
      "grad_norm": 0.09902913123369217,
      "learning_rate": 9.308460481992216e-05,
      "loss": 1.3726,
      "mean_token_accuracy": 0.7464862167835236,
      "num_tokens": 834395460.0,
      "step": 11610
    },
    {
      "entropy": 0.9381388798356056,
      "epoch": 3.1670754974107385,
      "grad_norm": 0.10641501098871231,
      "learning_rate": 9.306935171367086e-05,
      "loss": 1.3662,
      "mean_token_accuracy": 0.7477539911866188,
      "num_tokens": 835116330.0,
      "step": 11620
    },
    {
      "entropy": 0.9438267111778259,
      "epoch": 3.169801035704552,
      "grad_norm": 0.09885408729314804,
      "learning_rate": 9.30540830568225e-05,
      "loss": 1.3716,
      "mean_token_accuracy": 0.7419815719127655,
      "num_tokens": 835823829.0,
      "step": 11630
    },
    {
      "entropy": 0.9350500702857971,
      "epoch": 3.172526573998365,
      "grad_norm": 0.09166236221790314,
      "learning_rate": 9.303879885488998e-05,
      "loss": 1.3672,
      "mean_token_accuracy": 0.7425052419304847,
      "num_tokens": 836526949.0,
      "step": 11640
    },
    {
      "entropy": 0.939009566605091,
      "epoch": 3.1752521122921777,
      "grad_norm": 0.09826038032770157,
      "learning_rate": 9.302349911339179e-05,
      "loss": 1.3718,
      "mean_token_accuracy": 0.7404588326811791,
      "num_tokens": 837230575.0,
      "step": 11650
    },
    {
      "entropy": 0.9326612934470176,
      "epoch": 3.1779776505859907,
      "grad_norm": 0.09045529365539551,
      "learning_rate": 9.300818383785207e-05,
      "loss": 1.364,
      "mean_token_accuracy": 0.7500249862670898,
      "num_tokens": 837960290.0,
      "step": 11660
    },
    {
      "entropy": 0.9376754447817802,
      "epoch": 3.1807031888798036,
      "grad_norm": 0.10200121253728867,
      "learning_rate": 9.299285303380054e-05,
      "loss": 1.3701,
      "mean_token_accuracy": 0.7456318408250808,
      "num_tokens": 838683179.0,
      "step": 11670
    },
    {
      "entropy": 0.9378181710839272,
      "epoch": 3.1834287271736166,
      "grad_norm": 0.09331372380256653,
      "learning_rate": 9.297750670677252e-05,
      "loss": 1.3738,
      "mean_token_accuracy": 0.7449145019054413,
      "num_tokens": 839397651.0,
      "step": 11680
    },
    {
      "entropy": 0.9413901746273041,
      "epoch": 3.18615426546743,
      "grad_norm": 0.09418069571256638,
      "learning_rate": 9.296214486230894e-05,
      "loss": 1.3697,
      "mean_token_accuracy": 0.7454806089401245,
      "num_tokens": 840122308.0,
      "step": 11690
    },
    {
      "entropy": 0.9366207778453827,
      "epoch": 3.188879803761243,
      "grad_norm": 0.10092388093471527,
      "learning_rate": 9.294676750595637e-05,
      "loss": 1.3642,
      "mean_token_accuracy": 0.7400210842490196,
      "num_tokens": 840822855.0,
      "step": 11700
    },
    {
      "entropy": 0.9360883101820946,
      "epoch": 3.191605342055056,
      "grad_norm": 0.09489568322896957,
      "learning_rate": 9.293137464326694e-05,
      "loss": 1.3606,
      "mean_token_accuracy": 0.7458364591002464,
      "num_tokens": 841552929.0,
      "step": 11710
    },
    {
      "entropy": 0.9407029107213021,
      "epoch": 3.194330880348869,
      "grad_norm": 0.1279614418745041,
      "learning_rate": 9.291596627979836e-05,
      "loss": 1.3689,
      "mean_token_accuracy": 0.7444825232028961,
      "num_tokens": 842270506.0,
      "step": 11720
    },
    {
      "entropy": 0.933660639822483,
      "epoch": 3.1970564186426818,
      "grad_norm": 0.11487264931201935,
      "learning_rate": 9.290054242111399e-05,
      "loss": 1.3626,
      "mean_token_accuracy": 0.745450496673584,
      "num_tokens": 842997926.0,
      "step": 11730
    },
    {
      "entropy": 0.9356621518731117,
      "epoch": 3.199781956936495,
      "grad_norm": 0.09398036450147629,
      "learning_rate": 9.288510307278277e-05,
      "loss": 1.3693,
      "mean_token_accuracy": 0.7445752680301666,
      "num_tokens": 843719586.0,
      "step": 11740
    },
    {
      "entropy": 0.9460142493247986,
      "epoch": 3.202507495230308,
      "grad_norm": 0.11102558672428131,
      "learning_rate": 9.28696482403792e-05,
      "loss": 1.3744,
      "mean_token_accuracy": 0.743507881462574,
      "num_tokens": 844445492.0,
      "step": 11750
    },
    {
      "entropy": 0.9449904471635818,
      "epoch": 3.205233033524121,
      "grad_norm": 0.1021118238568306,
      "learning_rate": 9.28541779294834e-05,
      "loss": 1.3837,
      "mean_token_accuracy": 0.7422887161374092,
      "num_tokens": 845162237.0,
      "step": 11760
    },
    {
      "entropy": 0.9379758894443512,
      "epoch": 3.207958571817934,
      "grad_norm": 0.08937733620405197,
      "learning_rate": 9.283869214568108e-05,
      "loss": 1.3698,
      "mean_token_accuracy": 0.7471628934144974,
      "num_tokens": 845891839.0,
      "step": 11770
    },
    {
      "entropy": 0.9340906962752342,
      "epoch": 3.210684110111747,
      "grad_norm": 0.11052072793245316,
      "learning_rate": 9.282319089456353e-05,
      "loss": 1.3802,
      "mean_token_accuracy": 0.7459457963705063,
      "num_tokens": 846594518.0,
      "step": 11780
    },
    {
      "entropy": 0.9423430398106575,
      "epoch": 3.21340964840556,
      "grad_norm": 0.09794645756483078,
      "learning_rate": 9.280767418172762e-05,
      "loss": 1.3678,
      "mean_token_accuracy": 0.741601151227951,
      "num_tokens": 847305573.0,
      "step": 11790
    },
    {
      "entropy": 0.9387161493301391,
      "epoch": 3.2161351866993733,
      "grad_norm": 0.1304703652858734,
      "learning_rate": 9.27921420127758e-05,
      "loss": 1.3705,
      "mean_token_accuracy": 0.7438861384987832,
      "num_tokens": 848019099.0,
      "step": 11800
    },
    {
      "entropy": 0.9368794485926628,
      "epoch": 3.218860724993186,
      "grad_norm": 0.10444916784763336,
      "learning_rate": 9.27765943933161e-05,
      "loss": 1.3731,
      "mean_token_accuracy": 0.7477671071887017,
      "num_tokens": 848736852.0,
      "step": 11810
    },
    {
      "entropy": 0.9384323671460152,
      "epoch": 3.221586263286999,
      "grad_norm": 0.10987184941768646,
      "learning_rate": 9.276103132896214e-05,
      "loss": 1.3652,
      "mean_token_accuracy": 0.7445508196949959,
      "num_tokens": 849453291.0,
      "step": 11820
    },
    {
      "entropy": 0.9372527226805687,
      "epoch": 3.224311801580812,
      "grad_norm": 0.09234461933374405,
      "learning_rate": 9.274545282533313e-05,
      "loss": 1.3728,
      "mean_token_accuracy": 0.744299691915512,
      "num_tokens": 850168016.0,
      "step": 11830
    },
    {
      "entropy": 0.9347311824560165,
      "epoch": 3.227037339874625,
      "grad_norm": 0.08843154460191727,
      "learning_rate": 9.272985888805383e-05,
      "loss": 1.3616,
      "mean_token_accuracy": 0.7469219848513603,
      "num_tokens": 850885861.0,
      "step": 11840
    },
    {
      "entropy": 0.9396131202578545,
      "epoch": 3.2297628781684384,
      "grad_norm": 0.09732121229171753,
      "learning_rate": 9.271424952275456e-05,
      "loss": 1.3679,
      "mean_token_accuracy": 0.748391892015934,
      "num_tokens": 851615694.0,
      "step": 11850
    },
    {
      "entropy": 0.946087296307087,
      "epoch": 3.2324884164622514,
      "grad_norm": 0.14983533322811127,
      "learning_rate": 9.269862473507125e-05,
      "loss": 1.3732,
      "mean_token_accuracy": 0.7415955275297165,
      "num_tokens": 852338796.0,
      "step": 11860
    },
    {
      "entropy": 0.9348750323057174,
      "epoch": 3.2352139547560643,
      "grad_norm": 0.11568590998649597,
      "learning_rate": 9.268298453064537e-05,
      "loss": 1.3665,
      "mean_token_accuracy": 0.7486034065485001,
      "num_tokens": 853061523.0,
      "step": 11870
    },
    {
      "entropy": 0.9446531474590302,
      "epoch": 3.2379394930498773,
      "grad_norm": 0.15574434399604797,
      "learning_rate": 9.266732891512395e-05,
      "loss": 1.372,
      "mean_token_accuracy": 0.744923149049282,
      "num_tokens": 853778037.0,
      "step": 11880
    },
    {
      "entropy": 0.9377698779106141,
      "epoch": 3.24066503134369,
      "grad_norm": 0.09689576178789139,
      "learning_rate": 9.265165789415962e-05,
      "loss": 1.3722,
      "mean_token_accuracy": 0.7483894094824791,
      "num_tokens": 854501956.0,
      "step": 11890
    },
    {
      "entropy": 0.9404146298766136,
      "epoch": 3.2433905696375036,
      "grad_norm": 0.1202041283249855,
      "learning_rate": 9.263597147341052e-05,
      "loss": 1.3705,
      "mean_token_accuracy": 0.7417228803038597,
      "num_tokens": 855210524.0,
      "step": 11900
    },
    {
      "entropy": 0.9327780216932297,
      "epoch": 3.2461161079313166,
      "grad_norm": 0.09388814121484756,
      "learning_rate": 9.262026965854043e-05,
      "loss": 1.3603,
      "mean_token_accuracy": 0.7462678134441376,
      "num_tokens": 855938036.0,
      "step": 11910
    },
    {
      "entropy": 0.9489069446921349,
      "epoch": 3.2488416462251295,
      "grad_norm": 0.10653320699930191,
      "learning_rate": 9.26045524552186e-05,
      "loss": 1.374,
      "mean_token_accuracy": 0.7437566295266151,
      "num_tokens": 856669630.0,
      "step": 11920
    },
    {
      "entropy": 0.9383052349090576,
      "epoch": 3.2515671845189424,
      "grad_norm": 0.13587507605552673,
      "learning_rate": 9.258881986911988e-05,
      "loss": 1.373,
      "mean_token_accuracy": 0.7436003983020782,
      "num_tokens": 857384372.0,
      "step": 11930
    },
    {
      "entropy": 0.9415795043110847,
      "epoch": 3.2542927228127554,
      "grad_norm": 0.09867548942565918,
      "learning_rate": 9.257307190592468e-05,
      "loss": 1.3788,
      "mean_token_accuracy": 0.740178307890892,
      "num_tokens": 858089388.0,
      "step": 11940
    },
    {
      "entropy": 0.9481500625610352,
      "epoch": 3.257018261106569,
      "grad_norm": 0.2128295749425888,
      "learning_rate": 9.255730857131894e-05,
      "loss": 1.3832,
      "mean_token_accuracy": 0.7374232366681099,
      "num_tokens": 858803752.0,
      "step": 11950
    },
    {
      "entropy": 0.9385175615549087,
      "epoch": 3.2597437994003817,
      "grad_norm": 0.10609021782875061,
      "learning_rate": 9.254152987099416e-05,
      "loss": 1.3639,
      "mean_token_accuracy": 0.7469068959355354,
      "num_tokens": 859535472.0,
      "step": 11960
    },
    {
      "entropy": 0.9444188758730888,
      "epoch": 3.2624693376941947,
      "grad_norm": 0.14103463292121887,
      "learning_rate": 9.252573581064741e-05,
      "loss": 1.3789,
      "mean_token_accuracy": 0.7409269601106644,
      "num_tokens": 860247498.0,
      "step": 11970
    },
    {
      "entropy": 0.939168532192707,
      "epoch": 3.2651948759880076,
      "grad_norm": 0.09561815857887268,
      "learning_rate": 9.250992639598127e-05,
      "loss": 1.3722,
      "mean_token_accuracy": 0.7489908814430237,
      "num_tokens": 860973964.0,
      "step": 11980
    },
    {
      "entropy": 0.9405602127313614,
      "epoch": 3.2679204142818206,
      "grad_norm": 0.1375446617603302,
      "learning_rate": 9.249410163270388e-05,
      "loss": 1.3746,
      "mean_token_accuracy": 0.7432276397943497,
      "num_tokens": 861687960.0,
      "step": 11990
    },
    {
      "entropy": 0.9345800951123238,
      "epoch": 3.2706459525756335,
      "grad_norm": 0.09133090078830719,
      "learning_rate": 9.247826152652891e-05,
      "loss": 1.3644,
      "mean_token_accuracy": 0.7489025801420212,
      "num_tokens": 862417292.0,
      "step": 12000
    },
    {
      "entropy": 0.9332168892025947,
      "epoch": 3.273371490869447,
      "grad_norm": 0.10458455979824066,
      "learning_rate": 9.24624060831756e-05,
      "loss": 1.3626,
      "mean_token_accuracy": 0.7467944294214248,
      "num_tokens": 863144584.0,
      "step": 12010
    },
    {
      "entropy": 0.939929710328579,
      "epoch": 3.27609702916326,
      "grad_norm": 0.09958841651678085,
      "learning_rate": 9.244653530836869e-05,
      "loss": 1.3687,
      "mean_token_accuracy": 0.7472238644957543,
      "num_tokens": 863874990.0,
      "step": 12020
    },
    {
      "entropy": 0.935688091814518,
      "epoch": 3.278822567457073,
      "grad_norm": 0.10171917825937271,
      "learning_rate": 9.243064920783848e-05,
      "loss": 1.3653,
      "mean_token_accuracy": 0.7491013899445533,
      "num_tokens": 864608673.0,
      "step": 12030
    },
    {
      "entropy": 0.9424743622541427,
      "epoch": 3.2815481057508857,
      "grad_norm": 0.1102117970585823,
      "learning_rate": 9.241474778732082e-05,
      "loss": 1.3759,
      "mean_token_accuracy": 0.745431300997734,
      "num_tokens": 865331852.0,
      "step": 12040
    },
    {
      "entropy": 0.9367278933525085,
      "epoch": 3.2842736440446987,
      "grad_norm": 0.10428949445486069,
      "learning_rate": 9.239883105255702e-05,
      "loss": 1.3631,
      "mean_token_accuracy": 0.7427160575985908,
      "num_tokens": 866052048.0,
      "step": 12050
    },
    {
      "entropy": 0.9408783093094826,
      "epoch": 3.286999182338512,
      "grad_norm": 0.08694595843553543,
      "learning_rate": 9.238289900929402e-05,
      "loss": 1.3697,
      "mean_token_accuracy": 0.7473935082554817,
      "num_tokens": 866783724.0,
      "step": 12060
    },
    {
      "entropy": 0.936025595664978,
      "epoch": 3.289724720632325,
      "grad_norm": 0.12551167607307434,
      "learning_rate": 9.236695166328419e-05,
      "loss": 1.3586,
      "mean_token_accuracy": 0.7464142009615898,
      "num_tokens": 867514322.0,
      "step": 12070
    },
    {
      "entropy": 0.9376924529671669,
      "epoch": 3.292450258926138,
      "grad_norm": 0.10297540575265884,
      "learning_rate": 9.235098902028549e-05,
      "loss": 1.3668,
      "mean_token_accuracy": 0.7453285813331604,
      "num_tokens": 868233798.0,
      "step": 12080
    },
    {
      "entropy": 0.9396264761686325,
      "epoch": 3.295175797219951,
      "grad_norm": 0.11221079528331757,
      "learning_rate": 9.233501108606139e-05,
      "loss": 1.3788,
      "mean_token_accuracy": 0.7407119363546372,
      "num_tokens": 868931400.0,
      "step": 12090
    },
    {
      "entropy": 0.9314264595508576,
      "epoch": 3.297901335513764,
      "grad_norm": 0.12180259823799133,
      "learning_rate": 9.231901786638087e-05,
      "loss": 1.3588,
      "mean_token_accuracy": 0.7533206358551979,
      "num_tokens": 869673649.0,
      "step": 12100
    },
    {
      "entropy": 0.9336747452616692,
      "epoch": 3.300626873807577,
      "grad_norm": 0.0984652042388916,
      "learning_rate": 9.230300936701843e-05,
      "loss": 1.3633,
      "mean_token_accuracy": 0.7461344450712204,
      "num_tokens": 870392970.0,
      "step": 12110
    },
    {
      "entropy": 0.9405874907970428,
      "epoch": 3.30335241210139,
      "grad_norm": 0.10068907588720322,
      "learning_rate": 9.228698559375409e-05,
      "loss": 1.3769,
      "mean_token_accuracy": 0.7446421384811401,
      "num_tokens": 871119240.0,
      "step": 12120
    },
    {
      "entropy": 0.9353307262063026,
      "epoch": 3.306077950395203,
      "grad_norm": 0.08579584211111069,
      "learning_rate": 9.227094655237338e-05,
      "loss": 1.3713,
      "mean_token_accuracy": 0.7481286942958831,
      "num_tokens": 871843150.0,
      "step": 12130
    },
    {
      "entropy": 0.9327404066920281,
      "epoch": 3.308803488689016,
      "grad_norm": 0.09677369147539139,
      "learning_rate": 9.225489224866737e-05,
      "loss": 1.3648,
      "mean_token_accuracy": 0.7528234973549843,
      "num_tokens": 872575930.0,
      "step": 12140
    },
    {
      "entropy": 0.9381903916597366,
      "epoch": 3.311529026982829,
      "grad_norm": 0.11243049800395966,
      "learning_rate": 9.223882268843259e-05,
      "loss": 1.3757,
      "mean_token_accuracy": 0.7431700929999352,
      "num_tokens": 873293545.0,
      "step": 12150
    },
    {
      "entropy": 0.9382534325122833,
      "epoch": 3.314254565276642,
      "grad_norm": 0.08840969949960709,
      "learning_rate": 9.222273787747112e-05,
      "loss": 1.366,
      "mean_token_accuracy": 0.7444361537694931,
      "num_tokens": 874014084.0,
      "step": 12160
    },
    {
      "entropy": 0.9425742998719215,
      "epoch": 3.3169801035704554,
      "grad_norm": 0.11762413382530212,
      "learning_rate": 9.220663782159054e-05,
      "loss": 1.3804,
      "mean_token_accuracy": 0.7423619404435158,
      "num_tokens": 874719773.0,
      "step": 12170
    },
    {
      "entropy": 0.9365145936608315,
      "epoch": 3.3197056418642683,
      "grad_norm": 0.11608080565929413,
      "learning_rate": 9.219052252660392e-05,
      "loss": 1.3688,
      "mean_token_accuracy": 0.7462546646595001,
      "num_tokens": 875430194.0,
      "step": 12180
    },
    {
      "entropy": 0.9391071319580078,
      "epoch": 3.3224311801580813,
      "grad_norm": 0.09551326930522919,
      "learning_rate": 9.217439199832982e-05,
      "loss": 1.3701,
      "mean_token_accuracy": 0.7475197538733482,
      "num_tokens": 876156104.0,
      "step": 12190
    },
    {
      "entropy": 0.9328495651483536,
      "epoch": 3.325156718451894,
      "grad_norm": 0.10556455701589584,
      "learning_rate": 9.215824624259235e-05,
      "loss": 1.3674,
      "mean_token_accuracy": 0.7482963696122169,
      "num_tokens": 876877994.0,
      "step": 12200
    },
    {
      "entropy": 0.936648765206337,
      "epoch": 3.327882256745707,
      "grad_norm": 0.09045548737049103,
      "learning_rate": 9.214208526522109e-05,
      "loss": 1.3608,
      "mean_token_accuracy": 0.746910671889782,
      "num_tokens": 877606048.0,
      "step": 12210
    },
    {
      "entropy": 0.934162038564682,
      "epoch": 3.33060779503952,
      "grad_norm": 0.10036295652389526,
      "learning_rate": 9.212590907205108e-05,
      "loss": 1.3626,
      "mean_token_accuracy": 0.7505731761455536,
      "num_tokens": 878339534.0,
      "step": 12220
    },
    {
      "entropy": 0.9365512371063233,
      "epoch": 3.3333333333333335,
      "grad_norm": 0.15659716725349426,
      "learning_rate": 9.210971766892291e-05,
      "loss": 1.3748,
      "mean_token_accuracy": 0.7418654069304467,
      "num_tokens": 879043376.0,
      "step": 12230
    },
    {
      "entropy": 0.9365818828344346,
      "epoch": 3.3360588716271464,
      "grad_norm": 0.10925443470478058,
      "learning_rate": 9.209351106168265e-05,
      "loss": 1.369,
      "mean_token_accuracy": 0.7447827830910683,
      "num_tokens": 879761153.0,
      "step": 12240
    },
    {
      "entropy": 0.9344280377030373,
      "epoch": 3.3387844099209594,
      "grad_norm": 0.1060609519481659,
      "learning_rate": 9.207728925618184e-05,
      "loss": 1.3692,
      "mean_token_accuracy": 0.7418735533952713,
      "num_tokens": 880472362.0,
      "step": 12250
    },
    {
      "entropy": 0.9354407399892807,
      "epoch": 3.3415099482147723,
      "grad_norm": 0.09413611888885498,
      "learning_rate": 9.206105225827751e-05,
      "loss": 1.3643,
      "mean_token_accuracy": 0.7522047266364098,
      "num_tokens": 881229725.0,
      "step": 12260
    },
    {
      "entropy": 0.9443309381604195,
      "epoch": 3.3442354865085853,
      "grad_norm": 0.09533295035362244,
      "learning_rate": 9.204480007383218e-05,
      "loss": 1.3746,
      "mean_token_accuracy": 0.7469098627567291,
      "num_tokens": 881961980.0,
      "step": 12270
    },
    {
      "entropy": 0.9417243987321854,
      "epoch": 3.3469610248023987,
      "grad_norm": 0.11531821638345718,
      "learning_rate": 9.202853270871387e-05,
      "loss": 1.3679,
      "mean_token_accuracy": 0.7430923923850059,
      "num_tokens": 882679094.0,
      "step": 12280
    },
    {
      "entropy": 0.9415391772985459,
      "epoch": 3.3496865630962116,
      "grad_norm": 0.09958190470933914,
      "learning_rate": 9.201225016879604e-05,
      "loss": 1.38,
      "mean_token_accuracy": 0.74543047696352,
      "num_tokens": 883399861.0,
      "step": 12290
    },
    {
      "entropy": 0.9387096092104912,
      "epoch": 3.3524121013900245,
      "grad_norm": 0.1257648766040802,
      "learning_rate": 9.19959524599577e-05,
      "loss": 1.3643,
      "mean_token_accuracy": 0.7435487121343612,
      "num_tokens": 884116502.0,
      "step": 12300
    },
    {
      "entropy": 0.941948288679123,
      "epoch": 3.3551376396838375,
      "grad_norm": 0.11715131253004074,
      "learning_rate": 9.197963958808323e-05,
      "loss": 1.3747,
      "mean_token_accuracy": 0.7438194438815117,
      "num_tokens": 884835006.0,
      "step": 12310
    },
    {
      "entropy": 0.9320042371749878,
      "epoch": 3.3578631779776504,
      "grad_norm": 0.09540057182312012,
      "learning_rate": 9.19633115590626e-05,
      "loss": 1.3681,
      "mean_token_accuracy": 0.7444098517298698,
      "num_tokens": 885535322.0,
      "step": 12320
    },
    {
      "entropy": 0.9351300984621048,
      "epoch": 3.3605887162714634,
      "grad_norm": 0.10215869545936584,
      "learning_rate": 9.194696837879118e-05,
      "loss": 1.377,
      "mean_token_accuracy": 0.7417103961110115,
      "num_tokens": 886243121.0,
      "step": 12330
    },
    {
      "entropy": 0.9415004014968872,
      "epoch": 3.3633142545652768,
      "grad_norm": 0.09734909236431122,
      "learning_rate": 9.193061005316981e-05,
      "loss": 1.374,
      "mean_token_accuracy": 0.7447368308901787,
      "num_tokens": 886952902.0,
      "step": 12340
    },
    {
      "entropy": 0.9400848492980003,
      "epoch": 3.3660397928590897,
      "grad_norm": 0.10203968733549118,
      "learning_rate": 9.191423658810486e-05,
      "loss": 1.3722,
      "mean_token_accuracy": 0.7438996389508248,
      "num_tokens": 887663324.0,
      "step": 12350
    },
    {
      "entropy": 0.9361301407217979,
      "epoch": 3.3687653311529027,
      "grad_norm": 0.09793830662965775,
      "learning_rate": 9.189784798950809e-05,
      "loss": 1.3608,
      "mean_token_accuracy": 0.7440217867493629,
      "num_tokens": 888370938.0,
      "step": 12360
    },
    {
      "entropy": 0.9428472727537155,
      "epoch": 3.3714908694467156,
      "grad_norm": 0.09535105526447296,
      "learning_rate": 9.188144426329679e-05,
      "loss": 1.3808,
      "mean_token_accuracy": 0.7373283475637435,
      "num_tokens": 889059887.0,
      "step": 12370
    },
    {
      "entropy": 0.9310889944434166,
      "epoch": 3.374216407740529,
      "grad_norm": 0.09437306970357895,
      "learning_rate": 9.186502541539364e-05,
      "loss": 1.3552,
      "mean_token_accuracy": 0.7492224603891373,
      "num_tokens": 889796454.0,
      "step": 12380
    },
    {
      "entropy": 0.9451030522584916,
      "epoch": 3.376941946034342,
      "grad_norm": 0.11006307601928711,
      "learning_rate": 9.184859145172685e-05,
      "loss": 1.3699,
      "mean_token_accuracy": 0.7429471015930176,
      "num_tokens": 890509477.0,
      "step": 12390
    },
    {
      "entropy": 0.940475769340992,
      "epoch": 3.379667484328155,
      "grad_norm": 0.09745632857084274,
      "learning_rate": 9.183214237823005e-05,
      "loss": 1.369,
      "mean_token_accuracy": 0.7450324624776841,
      "num_tokens": 891225518.0,
      "step": 12400
    },
    {
      "entropy": 0.9360862731933594,
      "epoch": 3.382393022621968,
      "grad_norm": 0.10466349124908447,
      "learning_rate": 9.181567820084234e-05,
      "loss": 1.3733,
      "mean_token_accuracy": 0.7460363939404487,
      "num_tokens": 891946748.0,
      "step": 12410
    },
    {
      "entropy": 0.942741434276104,
      "epoch": 3.385118560915781,
      "grad_norm": 0.10499139875173569,
      "learning_rate": 9.179919892550826e-05,
      "loss": 1.3775,
      "mean_token_accuracy": 0.74194056391716,
      "num_tokens": 892665500.0,
      "step": 12420
    },
    {
      "entropy": 0.9336649075150489,
      "epoch": 3.3878440992095937,
      "grad_norm": 0.0979248657822609,
      "learning_rate": 9.178270455817779e-05,
      "loss": 1.3603,
      "mean_token_accuracy": 0.7498516023159028,
      "num_tokens": 893401884.0,
      "step": 12430
    },
    {
      "entropy": 0.9364969998598098,
      "epoch": 3.3905696375034067,
      "grad_norm": 0.10366225987672806,
      "learning_rate": 9.176619510480642e-05,
      "loss": 1.3678,
      "mean_token_accuracy": 0.7454469427466393,
      "num_tokens": 894121883.0,
      "step": 12440
    },
    {
      "entropy": 0.9303735733032227,
      "epoch": 3.39329517579722,
      "grad_norm": 0.09294752031564713,
      "learning_rate": 9.174967057135501e-05,
      "loss": 1.3564,
      "mean_token_accuracy": 0.7494041442871093,
      "num_tokens": 894849468.0,
      "step": 12450
    },
    {
      "entropy": 0.932018369436264,
      "epoch": 3.396020714091033,
      "grad_norm": 0.10084271430969238,
      "learning_rate": 9.173313096378991e-05,
      "loss": 1.3702,
      "mean_token_accuracy": 0.7492056339979172,
      "num_tokens": 895566986.0,
      "step": 12460
    },
    {
      "entropy": 0.932258027791977,
      "epoch": 3.398746252384846,
      "grad_norm": 0.10121307522058487,
      "learning_rate": 9.17165762880829e-05,
      "loss": 1.3621,
      "mean_token_accuracy": 0.7474600434303283,
      "num_tokens": 896290691.0,
      "step": 12470
    },
    {
      "entropy": 0.9325224250555039,
      "epoch": 3.401471790678659,
      "grad_norm": 0.12568046152591705,
      "learning_rate": 9.170000655021121e-05,
      "loss": 1.3616,
      "mean_token_accuracy": 0.7490670397877693,
      "num_tokens": 897022185.0,
      "step": 12480
    },
    {
      "entropy": 0.9375859543681144,
      "epoch": 3.4041973289724723,
      "grad_norm": 0.0965314507484436,
      "learning_rate": 9.16834217561575e-05,
      "loss": 1.3769,
      "mean_token_accuracy": 0.7438808307051659,
      "num_tokens": 897741744.0,
      "step": 12490
    },
    {
      "entropy": 0.9329176500439644,
      "epoch": 3.4069228672662852,
      "grad_norm": 0.10663661360740662,
      "learning_rate": 9.166682191190986e-05,
      "loss": 1.3673,
      "mean_token_accuracy": 0.745447650551796,
      "num_tokens": 898456811.0,
      "step": 12500
    },
    {
      "entropy": 0.9322167053818703,
      "epoch": 3.409648405560098,
      "grad_norm": 0.10179069638252258,
      "learning_rate": 9.165020702346181e-05,
      "loss": 1.3663,
      "mean_token_accuracy": 0.745252838730812,
      "num_tokens": 899170032.0,
      "step": 12510
    },
    {
      "entropy": 0.9398092135787011,
      "epoch": 3.412373943853911,
      "grad_norm": 0.11868678778409958,
      "learning_rate": 9.163357709681235e-05,
      "loss": 1.3745,
      "mean_token_accuracy": 0.7410765290260315,
      "num_tokens": 899871591.0,
      "step": 12520
    },
    {
      "entropy": 0.9274461969733239,
      "epoch": 3.415099482147724,
      "grad_norm": 0.1143931970000267,
      "learning_rate": 9.161693213796583e-05,
      "loss": 1.3687,
      "mean_token_accuracy": 0.7493044048547745,
      "num_tokens": 900592996.0,
      "step": 12530
    },
    {
      "entropy": 0.9415442749857903,
      "epoch": 3.417825020441537,
      "grad_norm": 0.13553395867347717,
      "learning_rate": 9.160027215293208e-05,
      "loss": 1.372,
      "mean_token_accuracy": 0.7440562471747398,
      "num_tokens": 901302994.0,
      "step": 12540
    },
    {
      "entropy": 0.9388555884361267,
      "epoch": 3.4205505587353504,
      "grad_norm": 0.10603881627321243,
      "learning_rate": 9.158359714772637e-05,
      "loss": 1.3604,
      "mean_token_accuracy": 0.7457233399152756,
      "num_tokens": 902031397.0,
      "step": 12550
    },
    {
      "entropy": 0.9392382368445397,
      "epoch": 3.4232760970291634,
      "grad_norm": 0.09645843505859375,
      "learning_rate": 9.156690712836935e-05,
      "loss": 1.3786,
      "mean_token_accuracy": 0.7428862899541855,
      "num_tokens": 902736615.0,
      "step": 12560
    },
    {
      "entropy": 0.9384090662002563,
      "epoch": 3.4260016353229763,
      "grad_norm": 0.09202127903699875,
      "learning_rate": 9.155020210088711e-05,
      "loss": 1.367,
      "mean_token_accuracy": 0.7447908714413642,
      "num_tokens": 903456864.0,
      "step": 12570
    },
    {
      "entropy": 0.9301142379641533,
      "epoch": 3.4287271736167892,
      "grad_norm": 0.09689704328775406,
      "learning_rate": 9.153348207131115e-05,
      "loss": 1.3555,
      "mean_token_accuracy": 0.7492977052927017,
      "num_tokens": 904179710.0,
      "step": 12580
    },
    {
      "entropy": 0.9400479078292847,
      "epoch": 3.431452711910602,
      "grad_norm": 0.10009034723043442,
      "learning_rate": 9.15167470456784e-05,
      "loss": 1.3684,
      "mean_token_accuracy": 0.7443844601511955,
      "num_tokens": 904896321.0,
      "step": 12590
    },
    {
      "entropy": 0.9228950247168541,
      "epoch": 3.4341782502044156,
      "grad_norm": 0.09664631634950638,
      "learning_rate": 9.149999703003122e-05,
      "loss": 1.3513,
      "mean_token_accuracy": 0.7498526141047478,
      "num_tokens": 905619229.0,
      "step": 12600
    },
    {
      "entropy": 0.9317604154348373,
      "epoch": 3.4369037884982285,
      "grad_norm": 0.10724034905433655,
      "learning_rate": 9.148323203041736e-05,
      "loss": 1.3644,
      "mean_token_accuracy": 0.7476500362157822,
      "num_tokens": 906337838.0,
      "step": 12610
    },
    {
      "entropy": 0.9352495178580285,
      "epoch": 3.4396293267920415,
      "grad_norm": 0.08909676223993301,
      "learning_rate": 9.146645205288993e-05,
      "loss": 1.3701,
      "mean_token_accuracy": 0.7471271619200707,
      "num_tokens": 907062083.0,
      "step": 12620
    },
    {
      "entropy": 0.9326487988233566,
      "epoch": 3.4423548650858544,
      "grad_norm": 0.12836824357509613,
      "learning_rate": 9.144965710350756e-05,
      "loss": 1.3692,
      "mean_token_accuracy": 0.740940372645855,
      "num_tokens": 907756822.0,
      "step": 12630
    },
    {
      "entropy": 0.9410655170679092,
      "epoch": 3.4450804033796674,
      "grad_norm": 0.10784759372472763,
      "learning_rate": 9.14328471883342e-05,
      "loss": 1.362,
      "mean_token_accuracy": 0.7482161849737168,
      "num_tokens": 908488298.0,
      "step": 12640
    },
    {
      "entropy": 0.9392810389399529,
      "epoch": 3.4478059416734803,
      "grad_norm": 0.09571303427219391,
      "learning_rate": 9.141602231343922e-05,
      "loss": 1.3699,
      "mean_token_accuracy": 0.7492217555642128,
      "num_tokens": 909218497.0,
      "step": 12650
    },
    {
      "entropy": 0.9285607561469078,
      "epoch": 3.4505314799672937,
      "grad_norm": 0.09034627676010132,
      "learning_rate": 9.139918248489743e-05,
      "loss": 1.3608,
      "mean_token_accuracy": 0.7473503425717354,
      "num_tokens": 909943537.0,
      "step": 12660
    },
    {
      "entropy": 0.9412695676088333,
      "epoch": 3.4532570182611066,
      "grad_norm": 0.11949262022972107,
      "learning_rate": 9.138232770878898e-05,
      "loss": 1.3691,
      "mean_token_accuracy": 0.7450414091348648,
      "num_tokens": 910658971.0,
      "step": 12670
    },
    {
      "entropy": 0.9309820994734764,
      "epoch": 3.4559825565549196,
      "grad_norm": 0.09633857011795044,
      "learning_rate": 9.136545799119944e-05,
      "loss": 1.3564,
      "mean_token_accuracy": 0.7498602598905564,
      "num_tokens": 911387408.0,
      "step": 12680
    },
    {
      "entropy": 0.926887808740139,
      "epoch": 3.4587080948487325,
      "grad_norm": 0.09766601771116257,
      "learning_rate": 9.134857333821982e-05,
      "loss": 1.3693,
      "mean_token_accuracy": 0.7483025774359703,
      "num_tokens": 912102244.0,
      "step": 12690
    },
    {
      "entropy": 0.9397025138139725,
      "epoch": 3.4614336331425455,
      "grad_norm": 0.10852885246276855,
      "learning_rate": 9.133167375594647e-05,
      "loss": 1.3586,
      "mean_token_accuracy": 0.7443624779582023,
      "num_tokens": 912820952.0,
      "step": 12700
    },
    {
      "entropy": 0.9383267760276794,
      "epoch": 3.464159171436359,
      "grad_norm": 0.08960741013288498,
      "learning_rate": 9.131475925048112e-05,
      "loss": 1.3673,
      "mean_token_accuracy": 0.7449337512254715,
      "num_tokens": 913547512.0,
      "step": 12710
    },
    {
      "entropy": 0.936745609343052,
      "epoch": 3.466884709730172,
      "grad_norm": 0.09380835294723511,
      "learning_rate": 9.129782982793094e-05,
      "loss": 1.3694,
      "mean_token_accuracy": 0.7446971386671066,
      "num_tokens": 914260113.0,
      "step": 12720
    },
    {
      "entropy": 0.9339314952492714,
      "epoch": 3.4696102480239848,
      "grad_norm": 0.09022416919469833,
      "learning_rate": 9.128088549440844e-05,
      "loss": 1.3694,
      "mean_token_accuracy": 0.7450258255004882,
      "num_tokens": 914970537.0,
      "step": 12730
    },
    {
      "entropy": 0.9304733410477638,
      "epoch": 3.4723357863177977,
      "grad_norm": 0.10127075761556625,
      "learning_rate": 9.126392625603153e-05,
      "loss": 1.3581,
      "mean_token_accuracy": 0.748434054851532,
      "num_tokens": 915688174.0,
      "step": 12740
    },
    {
      "entropy": 0.9351843550801278,
      "epoch": 3.4750613246116107,
      "grad_norm": 0.09760762006044388,
      "learning_rate": 9.124695211892354e-05,
      "loss": 1.3637,
      "mean_token_accuracy": 0.7453558519482613,
      "num_tokens": 916404714.0,
      "step": 12750
    },
    {
      "entropy": 0.9345199972391128,
      "epoch": 3.4777868629054236,
      "grad_norm": 0.11228025704622269,
      "learning_rate": 9.122996308921309e-05,
      "loss": 1.3686,
      "mean_token_accuracy": 0.7460799053311348,
      "num_tokens": 917125300.0,
      "step": 12760
    },
    {
      "entropy": 0.9374775588512421,
      "epoch": 3.480512401199237,
      "grad_norm": 0.09692800045013428,
      "learning_rate": 9.121295917303425e-05,
      "loss": 1.3638,
      "mean_token_accuracy": 0.7440088778734207,
      "num_tokens": 917844430.0,
      "step": 12770
    },
    {
      "entropy": 0.9233436644077301,
      "epoch": 3.48323793949305,
      "grad_norm": 0.1387101709842682,
      "learning_rate": 9.119594037652648e-05,
      "loss": 1.3625,
      "mean_token_accuracy": 0.7465911999344825,
      "num_tokens": 918543912.0,
      "step": 12780
    },
    {
      "entropy": 0.9400423854589463,
      "epoch": 3.485963477786863,
      "grad_norm": 0.10187558084726334,
      "learning_rate": 9.117890670583453e-05,
      "loss": 1.3727,
      "mean_token_accuracy": 0.7459829822182655,
      "num_tokens": 919270358.0,
      "step": 12790
    },
    {
      "entropy": 0.9362019613385201,
      "epoch": 3.488689016080676,
      "grad_norm": 0.09143900871276855,
      "learning_rate": 9.116185816710859e-05,
      "loss": 1.3654,
      "mean_token_accuracy": 0.7460013523697853,
      "num_tokens": 919985986.0,
      "step": 12800
    },
    {
      "entropy": 0.9333808422088623,
      "epoch": 3.4914145543744888,
      "grad_norm": 0.16597706079483032,
      "learning_rate": 9.114479476650419e-05,
      "loss": 1.366,
      "mean_token_accuracy": 0.7485702648758888,
      "num_tokens": 920708294.0,
      "step": 12810
    },
    {
      "entropy": 0.9295020744204521,
      "epoch": 3.494140092668302,
      "grad_norm": 0.10717319697141647,
      "learning_rate": 9.112771651018222e-05,
      "loss": 1.3651,
      "mean_token_accuracy": 0.7482924610376358,
      "num_tokens": 921426520.0,
      "step": 12820
    },
    {
      "entropy": 0.9408106982707978,
      "epoch": 3.496865630962115,
      "grad_norm": 0.11571063101291656,
      "learning_rate": 9.111062340430898e-05,
      "loss": 1.3744,
      "mean_token_accuracy": 0.7441819131374359,
      "num_tokens": 922147846.0,
      "step": 12830
    },
    {
      "entropy": 0.93288584202528,
      "epoch": 3.499591169255928,
      "grad_norm": 0.09528737515211105,
      "learning_rate": 9.109351545505607e-05,
      "loss": 1.365,
      "mean_token_accuracy": 0.7429381534457207,
      "num_tokens": 922852434.0,
      "step": 12840
    },
    {
      "entropy": 0.9288337647914886,
      "epoch": 3.502316707549741,
      "grad_norm": 0.10257086157798767,
      "learning_rate": 9.107639266860048e-05,
      "loss": 1.3665,
      "mean_token_accuracy": 0.7479848623275757,
      "num_tokens": 923569534.0,
      "step": 12850
    },
    {
      "entropy": 0.9374867156147957,
      "epoch": 3.505042245843554,
      "grad_norm": 0.12302909046411514,
      "learning_rate": 9.105925505112459e-05,
      "loss": 1.3666,
      "mean_token_accuracy": 0.7516830086708068,
      "num_tokens": 924312732.0,
      "step": 12860
    },
    {
      "entropy": 0.9320302456617355,
      "epoch": 3.507767784137367,
      "grad_norm": 0.11141735315322876,
      "learning_rate": 9.104210260881603e-05,
      "loss": 1.3757,
      "mean_token_accuracy": 0.7412846013903618,
      "num_tokens": 925006320.0,
      "step": 12870
    },
    {
      "entropy": 0.9339482218027115,
      "epoch": 3.5104933224311803,
      "grad_norm": 0.13289707899093628,
      "learning_rate": 9.102493534786791e-05,
      "loss": 1.3694,
      "mean_token_accuracy": 0.7464481011033058,
      "num_tokens": 925723440.0,
      "step": 12880
    },
    {
      "entropy": 0.9333376854658126,
      "epoch": 3.5132188607249932,
      "grad_norm": 0.10286711156368256,
      "learning_rate": 9.10077532744786e-05,
      "loss": 1.3629,
      "mean_token_accuracy": 0.7458906590938568,
      "num_tokens": 926441778.0,
      "step": 12890
    },
    {
      "entropy": 0.9332393199205399,
      "epoch": 3.515944399018806,
      "grad_norm": 0.12085429579019547,
      "learning_rate": 9.099055639485187e-05,
      "loss": 1.368,
      "mean_token_accuracy": 0.7435436382889747,
      "num_tokens": 927157927.0,
      "step": 12900
    },
    {
      "entropy": 0.9383823603391648,
      "epoch": 3.518669937312619,
      "grad_norm": 0.09964218735694885,
      "learning_rate": 9.09733447151968e-05,
      "loss": 1.3766,
      "mean_token_accuracy": 0.7474200412631035,
      "num_tokens": 927874440.0,
      "step": 12910
    },
    {
      "entropy": 0.9362079545855522,
      "epoch": 3.5213954756064325,
      "grad_norm": 0.09498178213834763,
      "learning_rate": 9.095611824172785e-05,
      "loss": 1.3645,
      "mean_token_accuracy": 0.7498639911413193,
      "num_tokens": 928605569.0,
      "step": 12920
    },
    {
      "entropy": 0.926796767115593,
      "epoch": 3.5241210139002455,
      "grad_norm": 0.10993527621030807,
      "learning_rate": 9.09388769806648e-05,
      "loss": 1.3565,
      "mean_token_accuracy": 0.7513145238161087,
      "num_tokens": 929341369.0,
      "step": 12930
    },
    {
      "entropy": 0.9336249455809593,
      "epoch": 3.5268465521940584,
      "grad_norm": 0.09894877672195435,
      "learning_rate": 9.092162093823274e-05,
      "loss": 1.3683,
      "mean_token_accuracy": 0.7437681227922439,
      "num_tokens": 930057958.0,
      "step": 12940
    },
    {
      "entropy": 0.9399918019771576,
      "epoch": 3.5295720904878713,
      "grad_norm": 0.09787449240684509,
      "learning_rate": 9.090435012066214e-05,
      "loss": 1.3733,
      "mean_token_accuracy": 0.7445860520005226,
      "num_tokens": 930771529.0,
      "step": 12950
    },
    {
      "entropy": 0.9386146873235702,
      "epoch": 3.5322976287816843,
      "grad_norm": 0.11092506349086761,
      "learning_rate": 9.088706453418881e-05,
      "loss": 1.3753,
      "mean_token_accuracy": 0.7453641518950462,
      "num_tokens": 931482333.0,
      "step": 12960
    },
    {
      "entropy": 0.9318419069051742,
      "epoch": 3.5350231670754972,
      "grad_norm": 0.11447679996490479,
      "learning_rate": 9.086976418505386e-05,
      "loss": 1.3583,
      "mean_token_accuracy": 0.7467338860034942,
      "num_tokens": 932214332.0,
      "step": 12970
    },
    {
      "entropy": 0.9419029995799064,
      "epoch": 3.53774870536931,
      "grad_norm": 0.10076761245727539,
      "learning_rate": 9.085244907950377e-05,
      "loss": 1.3798,
      "mean_token_accuracy": 0.7389994010329246,
      "num_tokens": 932917557.0,
      "step": 12980
    },
    {
      "entropy": 0.9376365810632705,
      "epoch": 3.5404742436631236,
      "grad_norm": 0.1359153538942337,
      "learning_rate": 9.083511922379027e-05,
      "loss": 1.3728,
      "mean_token_accuracy": 0.7453759759664536,
      "num_tokens": 933632139.0,
      "step": 12990
    },
    {
      "entropy": 0.9354806005954742,
      "epoch": 3.5431997819569365,
      "grad_norm": 0.1105719730257988,
      "learning_rate": 9.081777462417052e-05,
      "loss": 1.3637,
      "mean_token_accuracy": 0.7441539481282234,
      "num_tokens": 934349014.0,
      "step": 13000
    },
    {
      "entropy": 0.9374251648783684,
      "epoch": 3.5459253202507495,
      "grad_norm": 0.12063316255807877,
      "learning_rate": 9.080041528690693e-05,
      "loss": 1.3637,
      "mean_token_accuracy": 0.7513807326555252,
      "num_tokens": 935097758.0,
      "step": 13010
    },
    {
      "entropy": 0.9375441312789917,
      "epoch": 3.5486508585445624,
      "grad_norm": 0.09859773516654968,
      "learning_rate": 9.078304121826724e-05,
      "loss": 1.3669,
      "mean_token_accuracy": 0.7448895171284675,
      "num_tokens": 935816503.0,
      "step": 13020
    },
    {
      "entropy": 0.9408703297376633,
      "epoch": 3.551376396838376,
      "grad_norm": 0.18847866356372833,
      "learning_rate": 9.076565242452455e-05,
      "loss": 1.3767,
      "mean_token_accuracy": 0.7430540844798088,
      "num_tokens": 936524438.0,
      "step": 13030
    },
    {
      "entropy": 0.9380682498216629,
      "epoch": 3.5541019351321887,
      "grad_norm": 0.11309590190649033,
      "learning_rate": 9.074824891195724e-05,
      "loss": 1.3647,
      "mean_token_accuracy": 0.7467644423246383,
      "num_tokens": 937248736.0,
      "step": 13040
    },
    {
      "entropy": 0.931859764456749,
      "epoch": 3.5568274734260017,
      "grad_norm": 0.09613960981369019,
      "learning_rate": 9.0730830686849e-05,
      "loss": 1.361,
      "mean_token_accuracy": 0.74162005931139,
      "num_tokens": 937957425.0,
      "step": 13050
    },
    {
      "entropy": 0.9350968182086945,
      "epoch": 3.5595530117198146,
      "grad_norm": 0.11296644061803818,
      "learning_rate": 9.071339775548886e-05,
      "loss": 1.3666,
      "mean_token_accuracy": 0.7488053023815155,
      "num_tokens": 938676202.0,
      "step": 13060
    },
    {
      "entropy": 0.9332686677575112,
      "epoch": 3.5622785500136276,
      "grad_norm": 0.10304231196641922,
      "learning_rate": 9.069595012417114e-05,
      "loss": 1.3547,
      "mean_token_accuracy": 0.7482245922088623,
      "num_tokens": 939400336.0,
      "step": 13070
    },
    {
      "entropy": 0.9414921462535858,
      "epoch": 3.5650040883074405,
      "grad_norm": 0.11928173899650574,
      "learning_rate": 9.067848779919546e-05,
      "loss": 1.376,
      "mean_token_accuracy": 0.7430342525243759,
      "num_tokens": 940115055.0,
      "step": 13080
    },
    {
      "entropy": 0.9392595738172531,
      "epoch": 3.5677296266012535,
      "grad_norm": 0.12385665625333786,
      "learning_rate": 9.066101078686677e-05,
      "loss": 1.3773,
      "mean_token_accuracy": 0.741777703166008,
      "num_tokens": 940818450.0,
      "step": 13090
    },
    {
      "entropy": 0.941134275496006,
      "epoch": 3.570455164895067,
      "grad_norm": 0.11523745208978653,
      "learning_rate": 9.064351909349531e-05,
      "loss": 1.3672,
      "mean_token_accuracy": 0.7461943909525871,
      "num_tokens": 941552084.0,
      "step": 13100
    },
    {
      "entropy": 0.9355636045336724,
      "epoch": 3.57318070318888,
      "grad_norm": 0.1081465482711792,
      "learning_rate": 9.062601272539662e-05,
      "loss": 1.3623,
      "mean_token_accuracy": 0.7465676367282867,
      "num_tokens": 942269624.0,
      "step": 13110
    },
    {
      "entropy": 0.9370079949498177,
      "epoch": 3.5759062414826928,
      "grad_norm": 0.10074573755264282,
      "learning_rate": 9.060849168889154e-05,
      "loss": 1.3791,
      "mean_token_accuracy": 0.7425153732299805,
      "num_tokens": 942969526.0,
      "step": 13120
    },
    {
      "entropy": 0.933417497575283,
      "epoch": 3.5786317797765057,
      "grad_norm": 0.1655840426683426,
      "learning_rate": 9.05909559903062e-05,
      "loss": 1.3626,
      "mean_token_accuracy": 0.7470695838332176,
      "num_tokens": 943692226.0,
      "step": 13130
    },
    {
      "entropy": 0.9250888034701348,
      "epoch": 3.581357318070319,
      "grad_norm": 0.13959792256355286,
      "learning_rate": 9.057340563597205e-05,
      "loss": 1.3594,
      "mean_token_accuracy": 0.7487790286540985,
      "num_tokens": 944415749.0,
      "step": 13140
    },
    {
      "entropy": 0.9316915094852447,
      "epoch": 3.584082856364132,
      "grad_norm": 0.1401204615831375,
      "learning_rate": 9.055584063222579e-05,
      "loss": 1.3663,
      "mean_token_accuracy": 0.7429968804121018,
      "num_tokens": 945119302.0,
      "step": 13150
    },
    {
      "entropy": 0.9314353331923485,
      "epoch": 3.586808394657945,
      "grad_norm": 0.09463392198085785,
      "learning_rate": 9.053826098540942e-05,
      "loss": 1.3652,
      "mean_token_accuracy": 0.7450727388262749,
      "num_tokens": 945824979.0,
      "step": 13160
    },
    {
      "entropy": 0.9322601348161698,
      "epoch": 3.589533932951758,
      "grad_norm": 0.10104291886091232,
      "learning_rate": 9.05206667018703e-05,
      "loss": 1.3625,
      "mean_token_accuracy": 0.7483246713876724,
      "num_tokens": 946550467.0,
      "step": 13170
    },
    {
      "entropy": 0.9309192970395088,
      "epoch": 3.592259471245571,
      "grad_norm": 0.10341756790876389,
      "learning_rate": 9.050305778796094e-05,
      "loss": 1.362,
      "mean_token_accuracy": 0.7465433076024055,
      "num_tokens": 947271473.0,
      "step": 13180
    },
    {
      "entropy": 0.9290094763040543,
      "epoch": 3.594985009539384,
      "grad_norm": 0.09247362613677979,
      "learning_rate": 9.048543425003923e-05,
      "loss": 1.3598,
      "mean_token_accuracy": 0.7471193477511406,
      "num_tokens": 947997230.0,
      "step": 13190
    },
    {
      "entropy": 0.9424288645386696,
      "epoch": 3.597710547833197,
      "grad_norm": 0.09236718714237213,
      "learning_rate": 9.046779609446834e-05,
      "loss": 1.3648,
      "mean_token_accuracy": 0.7454434633255005,
      "num_tokens": 948727995.0,
      "step": 13200
    },
    {
      "entropy": 0.9284640043973923,
      "epoch": 3.60043608612701,
      "grad_norm": 0.10414078086614609,
      "learning_rate": 9.045014332761665e-05,
      "loss": 1.3612,
      "mean_token_accuracy": 0.7465564653277397,
      "num_tokens": 949445614.0,
      "step": 13210
    },
    {
      "entropy": 0.9378997027873993,
      "epoch": 3.603161624420823,
      "grad_norm": 0.12277641892433167,
      "learning_rate": 9.043247595585789e-05,
      "loss": 1.3739,
      "mean_token_accuracy": 0.7436944678425789,
      "num_tokens": 950156224.0,
      "step": 13220
    },
    {
      "entropy": 0.9347283899784088,
      "epoch": 3.605887162714636,
      "grad_norm": 0.09808440506458282,
      "learning_rate": 9.041479398557102e-05,
      "loss": 1.363,
      "mean_token_accuracy": 0.7425682976841926,
      "num_tokens": 950874203.0,
      "step": 13230
    },
    {
      "entropy": 0.9326068714261055,
      "epoch": 3.6086127010084494,
      "grad_norm": 0.10203598439693451,
      "learning_rate": 9.03970974231403e-05,
      "loss": 1.357,
      "mean_token_accuracy": 0.7526075363159179,
      "num_tokens": 951618096.0,
      "step": 13240
    },
    {
      "entropy": 0.9280758410692215,
      "epoch": 3.6113382393022624,
      "grad_norm": 0.11678113788366318,
      "learning_rate": 9.037938627495521e-05,
      "loss": 1.3632,
      "mean_token_accuracy": 0.7487672224640847,
      "num_tokens": 952333770.0,
      "step": 13250
    },
    {
      "entropy": 0.9286215782165528,
      "epoch": 3.6140637775960753,
      "grad_norm": 0.107167087495327,
      "learning_rate": 9.036166054741055e-05,
      "loss": 1.3656,
      "mean_token_accuracy": 0.7437203407287598,
      "num_tokens": 953043793.0,
      "step": 13260
    },
    {
      "entropy": 0.9332890540361405,
      "epoch": 3.6167893158898883,
      "grad_norm": 0.08996494859457016,
      "learning_rate": 9.034392024690636e-05,
      "loss": 1.3652,
      "mean_token_accuracy": 0.7491186946630478,
      "num_tokens": 953771556.0,
      "step": 13270
    },
    {
      "entropy": 0.9364136576652526,
      "epoch": 3.6195148541837012,
      "grad_norm": 0.10558682680130005,
      "learning_rate": 9.032616537984792e-05,
      "loss": 1.3694,
      "mean_token_accuracy": 0.7433787032961845,
      "num_tokens": 954468628.0,
      "step": 13280
    },
    {
      "entropy": 0.9318516358733178,
      "epoch": 3.622240392477514,
      "grad_norm": 0.09815244376659393,
      "learning_rate": 9.030839595264584e-05,
      "loss": 1.365,
      "mean_token_accuracy": 0.7461149856448174,
      "num_tokens": 955179194.0,
      "step": 13290
    },
    {
      "entropy": 0.9292657166719437,
      "epoch": 3.624965930771327,
      "grad_norm": 0.09705313295125961,
      "learning_rate": 9.029061197171589e-05,
      "loss": 1.3623,
      "mean_token_accuracy": 0.7456430450081826,
      "num_tokens": 955893302.0,
      "step": 13300
    },
    {
      "entropy": 0.9309269085526466,
      "epoch": 3.6276914690651405,
      "grad_norm": 0.11863958090543747,
      "learning_rate": 9.027281344347918e-05,
      "loss": 1.3595,
      "mean_token_accuracy": 0.7468349114060402,
      "num_tokens": 956624858.0,
      "step": 13310
    },
    {
      "entropy": 0.934124930202961,
      "epoch": 3.6304170073589535,
      "grad_norm": 0.1297106146812439,
      "learning_rate": 9.025500037436202e-05,
      "loss": 1.368,
      "mean_token_accuracy": 0.7464649736881256,
      "num_tokens": 957347468.0,
      "step": 13320
    },
    {
      "entropy": 0.932029339671135,
      "epoch": 3.6331425456527664,
      "grad_norm": 0.10935456305742264,
      "learning_rate": 9.0237172770796e-05,
      "loss": 1.3577,
      "mean_token_accuracy": 0.7451001316308975,
      "num_tokens": 958056962.0,
      "step": 13330
    },
    {
      "entropy": 0.938096997141838,
      "epoch": 3.6358680839465793,
      "grad_norm": 0.15865057706832886,
      "learning_rate": 9.021933063921792e-05,
      "loss": 1.3714,
      "mean_token_accuracy": 0.7457519054412842,
      "num_tokens": 958779703.0,
      "step": 13340
    },
    {
      "entropy": 0.9339332118630409,
      "epoch": 3.6385936222403927,
      "grad_norm": 0.09937279671430588,
      "learning_rate": 9.020147398606987e-05,
      "loss": 1.3697,
      "mean_token_accuracy": 0.7467718929052353,
      "num_tokens": 959501476.0,
      "step": 13350
    },
    {
      "entropy": 0.9396980375051498,
      "epoch": 3.6413191605342057,
      "grad_norm": 0.11709866672754288,
      "learning_rate": 9.018360281779916e-05,
      "loss": 1.3692,
      "mean_token_accuracy": 0.7415199279785156,
      "num_tokens": 960208830.0,
      "step": 13360
    },
    {
      "entropy": 0.9336515188217163,
      "epoch": 3.6440446988280186,
      "grad_norm": 0.10379207134246826,
      "learning_rate": 9.016571714085835e-05,
      "loss": 1.36,
      "mean_token_accuracy": 0.7475280880928039,
      "num_tokens": 960936444.0,
      "step": 13370
    },
    {
      "entropy": 0.9317442283034325,
      "epoch": 3.6467702371218316,
      "grad_norm": 0.2132408767938614,
      "learning_rate": 9.014781696170522e-05,
      "loss": 1.3581,
      "mean_token_accuracy": 0.7497346103191376,
      "num_tokens": 961664188.0,
      "step": 13380
    },
    {
      "entropy": 0.9318334341049195,
      "epoch": 3.6494957754156445,
      "grad_norm": 0.09681301563978195,
      "learning_rate": 9.012990228680281e-05,
      "loss": 1.3658,
      "mean_token_accuracy": 0.7463125616312027,
      "num_tokens": 962383028.0,
      "step": 13390
    },
    {
      "entropy": 0.9316603437066078,
      "epoch": 3.6522213137094575,
      "grad_norm": 0.0922931507229805,
      "learning_rate": 9.011197312261938e-05,
      "loss": 1.3604,
      "mean_token_accuracy": 0.7499169021844864,
      "num_tokens": 963118555.0,
      "step": 13400
    },
    {
      "entropy": 0.9345208585262299,
      "epoch": 3.6549468520032704,
      "grad_norm": 0.10234454274177551,
      "learning_rate": 9.009402947562842e-05,
      "loss": 1.3572,
      "mean_token_accuracy": 0.7530842050909996,
      "num_tokens": 963860509.0,
      "step": 13410
    },
    {
      "entropy": 0.9333072051405906,
      "epoch": 3.657672390297084,
      "grad_norm": 0.26790082454681396,
      "learning_rate": 9.007607135230866e-05,
      "loss": 1.3628,
      "mean_token_accuracy": 0.7466099947690964,
      "num_tokens": 964578383.0,
      "step": 13420
    },
    {
      "entropy": 0.9337391629815102,
      "epoch": 3.6603979285908967,
      "grad_norm": 0.11218128353357315,
      "learning_rate": 9.005809875914406e-05,
      "loss": 1.3629,
      "mean_token_accuracy": 0.7469859942793846,
      "num_tokens": 965304208.0,
      "step": 13430
    },
    {
      "entropy": 0.9428641706705093,
      "epoch": 3.6631234668847097,
      "grad_norm": 0.10984963923692703,
      "learning_rate": 9.004011170262377e-05,
      "loss": 1.3796,
      "mean_token_accuracy": 0.7463513821363449,
      "num_tokens": 966032331.0,
      "step": 13440
    },
    {
      "entropy": 0.935543717443943,
      "epoch": 3.6658490051785226,
      "grad_norm": 0.1488730013370514,
      "learning_rate": 9.00221101892422e-05,
      "loss": 1.3682,
      "mean_token_accuracy": 0.7450001835823059,
      "num_tokens": 966741432.0,
      "step": 13450
    },
    {
      "entropy": 0.9354455277323723,
      "epoch": 3.668574543472336,
      "grad_norm": 0.09760552644729614,
      "learning_rate": 9.000409422549897e-05,
      "loss": 1.3604,
      "mean_token_accuracy": 0.7469463601708413,
      "num_tokens": 967463277.0,
      "step": 13460
    },
    {
      "entropy": 0.9351951852440834,
      "epoch": 3.671300081766149,
      "grad_norm": 0.09895797818899155,
      "learning_rate": 8.998606381789891e-05,
      "loss": 1.3709,
      "mean_token_accuracy": 0.745685450732708,
      "num_tokens": 968189722.0,
      "step": 13470
    },
    {
      "entropy": 0.9256167277693749,
      "epoch": 3.674025620059962,
      "grad_norm": 0.13834929466247559,
      "learning_rate": 8.996801897295206e-05,
      "loss": 1.3558,
      "mean_token_accuracy": 0.7506679490208625,
      "num_tokens": 968906248.0,
      "step": 13480
    },
    {
      "entropy": 0.933958075940609,
      "epoch": 3.676751158353775,
      "grad_norm": 0.12164489179849625,
      "learning_rate": 8.99499596971737e-05,
      "loss": 1.3721,
      "mean_token_accuracy": 0.7465602666139602,
      "num_tokens": 969633809.0,
      "step": 13490
    },
    {
      "entropy": 0.9461019173264503,
      "epoch": 3.679476696647588,
      "grad_norm": 0.1290566474199295,
      "learning_rate": 8.99318859970843e-05,
      "loss": 1.3738,
      "mean_token_accuracy": 0.7392740756273269,
      "num_tokens": 970329287.0,
      "step": 13500
    },
    {
      "entropy": 0.9286004945635795,
      "epoch": 3.6822022349414008,
      "grad_norm": 0.11958316713571548,
      "learning_rate": 8.991379787920952e-05,
      "loss": 1.3624,
      "mean_token_accuracy": 0.7454740762710571,
      "num_tokens": 971039204.0,
      "step": 13510
    },
    {
      "entropy": 0.9378550440073014,
      "epoch": 3.6849277732352137,
      "grad_norm": 0.12413039058446884,
      "learning_rate": 8.989569535008027e-05,
      "loss": 1.3705,
      "mean_token_accuracy": 0.7478660434484482,
      "num_tokens": 971767195.0,
      "step": 13520
    },
    {
      "entropy": 0.9414548948407173,
      "epoch": 3.687653311529027,
      "grad_norm": 0.1187831312417984,
      "learning_rate": 8.987757841623263e-05,
      "loss": 1.3679,
      "mean_token_accuracy": 0.7386318236589432,
      "num_tokens": 972477428.0,
      "step": 13530
    },
    {
      "entropy": 0.9377189055085182,
      "epoch": 3.69037884982284,
      "grad_norm": 0.10989218950271606,
      "learning_rate": 8.98594470842079e-05,
      "loss": 1.3717,
      "mean_token_accuracy": 0.7464066311717034,
      "num_tokens": 973195901.0,
      "step": 13540
    },
    {
      "entropy": 0.9250073865056038,
      "epoch": 3.693104388116653,
      "grad_norm": 0.11505521088838577,
      "learning_rate": 8.984130136055256e-05,
      "loss": 1.3665,
      "mean_token_accuracy": 0.7527343273162842,
      "num_tokens": 973919654.0,
      "step": 13550
    },
    {
      "entropy": 0.9378981202840805,
      "epoch": 3.695829926410466,
      "grad_norm": 0.10788934677839279,
      "learning_rate": 8.982314125181828e-05,
      "loss": 1.3651,
      "mean_token_accuracy": 0.7484021842479706,
      "num_tokens": 974654678.0,
      "step": 13560
    },
    {
      "entropy": 0.9401387721300125,
      "epoch": 3.6985554647042793,
      "grad_norm": 0.1131954938173294,
      "learning_rate": 8.980496676456196e-05,
      "loss": 1.3701,
      "mean_token_accuracy": 0.7417485862970352,
      "num_tokens": 975374744.0,
      "step": 13570
    },
    {
      "entropy": 0.926028548181057,
      "epoch": 3.7012810029980923,
      "grad_norm": 0.10438834875822067,
      "learning_rate": 8.978677790534566e-05,
      "loss": 1.3645,
      "mean_token_accuracy": 0.7438564658164978,
      "num_tokens": 976083178.0,
      "step": 13580
    },
    {
      "entropy": 0.9385993152856826,
      "epoch": 3.704006541291905,
      "grad_norm": 0.09293225407600403,
      "learning_rate": 8.976857468073664e-05,
      "loss": 1.3676,
      "mean_token_accuracy": 0.7446450099349022,
      "num_tokens": 976801711.0,
      "step": 13590
    },
    {
      "entropy": 0.9331994846463203,
      "epoch": 3.706732079585718,
      "grad_norm": 0.12272260338068008,
      "learning_rate": 8.975035709730737e-05,
      "loss": 1.3578,
      "mean_token_accuracy": 0.7492441236972809,
      "num_tokens": 977538379.0,
      "step": 13600
    },
    {
      "entropy": 0.9382228672504425,
      "epoch": 3.709457617879531,
      "grad_norm": 0.1115153357386589,
      "learning_rate": 8.973212516163545e-05,
      "loss": 1.3662,
      "mean_token_accuracy": 0.7477366283535958,
      "num_tokens": 978274991.0,
      "step": 13610
    },
    {
      "entropy": 0.9352905541658402,
      "epoch": 3.712183156173344,
      "grad_norm": 0.11782050132751465,
      "learning_rate": 8.971387888030371e-05,
      "loss": 1.3645,
      "mean_token_accuracy": 0.7457715913653373,
      "num_tokens": 978996771.0,
      "step": 13620
    },
    {
      "entropy": 0.9319099530577659,
      "epoch": 3.714908694467157,
      "grad_norm": 0.09972216933965683,
      "learning_rate": 8.969561825990013e-05,
      "loss": 1.3648,
      "mean_token_accuracy": 0.7467517778277397,
      "num_tokens": 979717813.0,
      "step": 13630
    },
    {
      "entropy": 0.9356176108121872,
      "epoch": 3.7176342327609704,
      "grad_norm": 0.10494368523359299,
      "learning_rate": 8.967734330701791e-05,
      "loss": 1.3601,
      "mean_token_accuracy": 0.7447285264730453,
      "num_tokens": 980438185.0,
      "step": 13640
    },
    {
      "entropy": 0.9377966582775116,
      "epoch": 3.7203597710547833,
      "grad_norm": 0.09992780536413193,
      "learning_rate": 8.965905402825536e-05,
      "loss": 1.3628,
      "mean_token_accuracy": 0.7456589475274086,
      "num_tokens": 981161128.0,
      "step": 13650
    },
    {
      "entropy": 0.9311361506581306,
      "epoch": 3.7230853093485963,
      "grad_norm": 0.12282440811395645,
      "learning_rate": 8.9640750430216e-05,
      "loss": 1.3618,
      "mean_token_accuracy": 0.7463863477110863,
      "num_tokens": 981881084.0,
      "step": 13660
    },
    {
      "entropy": 0.9404643461108207,
      "epoch": 3.725810847642409,
      "grad_norm": 0.0928298830986023,
      "learning_rate": 8.962243251950852e-05,
      "loss": 1.3635,
      "mean_token_accuracy": 0.7443462312221527,
      "num_tokens": 982606770.0,
      "step": 13670
    },
    {
      "entropy": 0.9373415976762771,
      "epoch": 3.7285363859362226,
      "grad_norm": 0.09752299636602402,
      "learning_rate": 8.960410030274681e-05,
      "loss": 1.3663,
      "mean_token_accuracy": 0.7402950346469879,
      "num_tokens": 983310616.0,
      "step": 13680
    },
    {
      "entropy": 0.939019364118576,
      "epoch": 3.7312619242300356,
      "grad_norm": 0.09725892543792725,
      "learning_rate": 8.958575378654984e-05,
      "loss": 1.3708,
      "mean_token_accuracy": 0.7425294920802117,
      "num_tokens": 984025069.0,
      "step": 13690
    },
    {
      "entropy": 0.9378419637680053,
      "epoch": 3.7339874625238485,
      "grad_norm": 0.1080869808793068,
      "learning_rate": 8.956739297754182e-05,
      "loss": 1.3785,
      "mean_token_accuracy": 0.7412189736962318,
      "num_tokens": 984727533.0,
      "step": 13700
    },
    {
      "entropy": 0.940848433971405,
      "epoch": 3.7367130008176614,
      "grad_norm": 0.0943247526884079,
      "learning_rate": 8.95490178823521e-05,
      "loss": 1.3672,
      "mean_token_accuracy": 0.7435140967369079,
      "num_tokens": 985456228.0,
      "step": 13710
    },
    {
      "entropy": 0.940029938519001,
      "epoch": 3.7394385391114744,
      "grad_norm": 0.09355151653289795,
      "learning_rate": 8.953062850761515e-05,
      "loss": 1.3708,
      "mean_token_accuracy": 0.7490209862589836,
      "num_tokens": 986183257.0,
      "step": 13720
    },
    {
      "entropy": 0.9465588822960853,
      "epoch": 3.7421640774052873,
      "grad_norm": 0.09960291534662247,
      "learning_rate": 8.951222485997066e-05,
      "loss": 1.3769,
      "mean_token_accuracy": 0.7437508583068848,
      "num_tokens": 986903930.0,
      "step": 13730
    },
    {
      "entropy": 0.9407746538519859,
      "epoch": 3.7448896156991007,
      "grad_norm": 0.09473884105682373,
      "learning_rate": 8.949380694606344e-05,
      "loss": 1.372,
      "mean_token_accuracy": 0.7417246505618096,
      "num_tokens": 987615662.0,
      "step": 13740
    },
    {
      "entropy": 0.938712403178215,
      "epoch": 3.7476151539929137,
      "grad_norm": 0.117513507604599,
      "learning_rate": 8.947537477254344e-05,
      "loss": 1.3651,
      "mean_token_accuracy": 0.7487455070018768,
      "num_tokens": 988348914.0,
      "step": 13750
    },
    {
      "entropy": 0.9397893026471138,
      "epoch": 3.7503406922867266,
      "grad_norm": 0.1013086810708046,
      "learning_rate": 8.945692834606576e-05,
      "loss": 1.3685,
      "mean_token_accuracy": 0.7428163930773735,
      "num_tokens": 989065170.0,
      "step": 13760
    },
    {
      "entropy": 0.9395933210849762,
      "epoch": 3.7530662305805396,
      "grad_norm": 0.10276360809803009,
      "learning_rate": 8.943846767329067e-05,
      "loss": 1.3702,
      "mean_token_accuracy": 0.7439565107226371,
      "num_tokens": 989782054.0,
      "step": 13770
    },
    {
      "entropy": 0.9382280990481376,
      "epoch": 3.755791768874353,
      "grad_norm": 0.09564198553562164,
      "learning_rate": 8.941999276088357e-05,
      "loss": 1.3649,
      "mean_token_accuracy": 0.7506726637482644,
      "num_tokens": 990523711.0,
      "step": 13780
    },
    {
      "entropy": 0.9305368557572364,
      "epoch": 3.758517307168166,
      "grad_norm": 0.09616661071777344,
      "learning_rate": 8.940150361551502e-05,
      "loss": 1.3535,
      "mean_token_accuracy": 0.7454472824931144,
      "num_tokens": 991238524.0,
      "step": 13790
    },
    {
      "entropy": 0.9342294216156006,
      "epoch": 3.761242845461979,
      "grad_norm": 0.09664113819599152,
      "learning_rate": 8.938300024386067e-05,
      "loss": 1.3686,
      "mean_token_accuracy": 0.7458544835448265,
      "num_tokens": 991959601.0,
      "step": 13800
    },
    {
      "entropy": 0.9341562122106553,
      "epoch": 3.763968383755792,
      "grad_norm": 0.09050773084163666,
      "learning_rate": 8.936448265260135e-05,
      "loss": 1.3635,
      "mean_token_accuracy": 0.7478775039315224,
      "num_tokens": 992690565.0,
      "step": 13810
    },
    {
      "entropy": 0.934185953438282,
      "epoch": 3.7666939220496047,
      "grad_norm": 0.09130620956420898,
      "learning_rate": 8.934595084842302e-05,
      "loss": 1.3651,
      "mean_token_accuracy": 0.746742396056652,
      "num_tokens": 993407699.0,
      "step": 13820
    },
    {
      "entropy": 0.9289571881294251,
      "epoch": 3.7694194603434177,
      "grad_norm": 0.11175952851772308,
      "learning_rate": 8.932740483801675e-05,
      "loss": 1.371,
      "mean_token_accuracy": 0.7462221279740333,
      "num_tokens": 994120106.0,
      "step": 13830
    },
    {
      "entropy": 0.9358081087470055,
      "epoch": 3.7721449986372306,
      "grad_norm": 0.11956287920475006,
      "learning_rate": 8.930884462807876e-05,
      "loss": 1.367,
      "mean_token_accuracy": 0.7460283309221267,
      "num_tokens": 994832393.0,
      "step": 13840
    },
    {
      "entropy": 0.9386581629514694,
      "epoch": 3.774870536931044,
      "grad_norm": 0.11503902822732925,
      "learning_rate": 8.92902702253104e-05,
      "loss": 1.3646,
      "mean_token_accuracy": 0.74471585303545,
      "num_tokens": 995559272.0,
      "step": 13850
    },
    {
      "entropy": 0.9225244998931885,
      "epoch": 3.777596075224857,
      "grad_norm": 0.09499411284923553,
      "learning_rate": 8.92716816364181e-05,
      "loss": 1.3568,
      "mean_token_accuracy": 0.752425593137741,
      "num_tokens": 996284236.0,
      "step": 13860
    },
    {
      "entropy": 0.9270935028791427,
      "epoch": 3.78032161351867,
      "grad_norm": 0.11209368705749512,
      "learning_rate": 8.925307886811348e-05,
      "loss": 1.3545,
      "mean_token_accuracy": 0.7520903870463371,
      "num_tokens": 997026655.0,
      "step": 13870
    },
    {
      "entropy": 0.9376581728458404,
      "epoch": 3.783047151812483,
      "grad_norm": 0.10614333301782608,
      "learning_rate": 8.923446192711323e-05,
      "loss": 1.3647,
      "mean_token_accuracy": 0.7438096687197685,
      "num_tokens": 997748050.0,
      "step": 13880
    },
    {
      "entropy": 0.9339435130357743,
      "epoch": 3.7857726901062962,
      "grad_norm": 0.11074935644865036,
      "learning_rate": 8.921583082013917e-05,
      "loss": 1.3723,
      "mean_token_accuracy": 0.7464756935834884,
      "num_tokens": 998459595.0,
      "step": 13890
    },
    {
      "entropy": 0.9401899024844169,
      "epoch": 3.788498228400109,
      "grad_norm": 0.11788266897201538,
      "learning_rate": 8.919718555391825e-05,
      "loss": 1.371,
      "mean_token_accuracy": 0.7440694138407707,
      "num_tokens": 999175251.0,
      "step": 13900
    },
    {
      "entropy": 0.9381234616041183,
      "epoch": 3.791223766693922,
      "grad_norm": 0.11436308175325394,
      "learning_rate": 8.917852613518252e-05,
      "loss": 1.3626,
      "mean_token_accuracy": 0.7457627654075623,
      "num_tokens": 999894200.0,
      "step": 13910
    },
    {
      "entropy": 0.9294078722596169,
      "epoch": 3.793949304987735,
      "grad_norm": 0.21421048045158386,
      "learning_rate": 8.915985257066912e-05,
      "loss": 1.3555,
      "mean_token_accuracy": 0.7505694150924682,
      "num_tokens": 1000616684.0,
      "step": 13920
    },
    {
      "entropy": 0.9338700622320175,
      "epoch": 3.796674843281548,
      "grad_norm": 0.10059638321399689,
      "learning_rate": 8.914116486712034e-05,
      "loss": 1.3789,
      "mean_token_accuracy": 0.7447562694549561,
      "num_tokens": 1001323188.0,
      "step": 13930
    },
    {
      "entropy": 0.9370212912559509,
      "epoch": 3.799400381575361,
      "grad_norm": 0.1040932834148407,
      "learning_rate": 8.912246303128353e-05,
      "loss": 1.371,
      "mean_token_accuracy": 0.7461760893464089,
      "num_tokens": 1002042069.0,
      "step": 13940
    },
    {
      "entropy": 0.9390333220362663,
      "epoch": 3.802125919869174,
      "grad_norm": 0.11725642532110214,
      "learning_rate": 8.910374706991119e-05,
      "loss": 1.3687,
      "mean_token_accuracy": 0.7445036932826042,
      "num_tokens": 1002760744.0,
      "step": 13950
    },
    {
      "entropy": 0.9299334555864334,
      "epoch": 3.8048514581629873,
      "grad_norm": 0.09295396506786346,
      "learning_rate": 8.908501698976087e-05,
      "loss": 1.3592,
      "mean_token_accuracy": 0.7487318173050881,
      "num_tokens": 1003497885.0,
      "step": 13960
    },
    {
      "entropy": 0.933466374874115,
      "epoch": 3.8075769964568003,
      "grad_norm": 0.09541443735361099,
      "learning_rate": 8.906627279759527e-05,
      "loss": 1.3712,
      "mean_token_accuracy": 0.7474435135722161,
      "num_tokens": 1004218492.0,
      "step": 13970
    },
    {
      "entropy": 0.9383569300174713,
      "epoch": 3.810302534750613,
      "grad_norm": 0.10156364738941193,
      "learning_rate": 8.904751450018213e-05,
      "loss": 1.3726,
      "mean_token_accuracy": 0.743280728161335,
      "num_tokens": 1004933440.0,
      "step": 13980
    },
    {
      "entropy": 0.9356719806790352,
      "epoch": 3.813028073044426,
      "grad_norm": 0.09579475224018097,
      "learning_rate": 8.902874210429433e-05,
      "loss": 1.3691,
      "mean_token_accuracy": 0.7394178152084351,
      "num_tokens": 1005627879.0,
      "step": 13990
    },
    {
      "entropy": 0.9362627252936363,
      "epoch": 3.8157536113382395,
      "grad_norm": 0.11506523191928864,
      "learning_rate": 8.900995561670982e-05,
      "loss": 1.3698,
      "mean_token_accuracy": 0.7435395658016205,
      "num_tokens": 1006338691.0,
      "step": 14000
    },
    {
      "entropy": 0.9338663190603256,
      "epoch": 3.8184791496320525,
      "grad_norm": 0.11216975748538971,
      "learning_rate": 8.899115504421163e-05,
      "loss": 1.3698,
      "mean_token_accuracy": 0.7448881000280381,
      "num_tokens": 1007054215.0,
      "step": 14010
    },
    {
      "entropy": 0.9339838832616806,
      "epoch": 3.8212046879258654,
      "grad_norm": 0.10040973871946335,
      "learning_rate": 8.897234039358788e-05,
      "loss": 1.369,
      "mean_token_accuracy": 0.7447676122188568,
      "num_tokens": 1007770956.0,
      "step": 14020
    },
    {
      "entropy": 0.9418065786361695,
      "epoch": 3.8239302262196784,
      "grad_norm": 0.13536158204078674,
      "learning_rate": 8.89535116716318e-05,
      "loss": 1.3724,
      "mean_token_accuracy": 0.7403697341680526,
      "num_tokens": 1008480927.0,
      "step": 14030
    },
    {
      "entropy": 0.9388413935899734,
      "epoch": 3.8266557645134913,
      "grad_norm": 0.11635293811559677,
      "learning_rate": 8.893466888514166e-05,
      "loss": 1.3625,
      "mean_token_accuracy": 0.7449570685625077,
      "num_tokens": 1009204090.0,
      "step": 14040
    },
    {
      "entropy": 0.9321177437901497,
      "epoch": 3.8293813028073043,
      "grad_norm": 0.09772279858589172,
      "learning_rate": 8.891581204092082e-05,
      "loss": 1.3646,
      "mean_token_accuracy": 0.7463148146867752,
      "num_tokens": 1009911598.0,
      "step": 14050
    },
    {
      "entropy": 0.9374329909682274,
      "epoch": 3.832106841101117,
      "grad_norm": 0.11084318906068802,
      "learning_rate": 8.889694114577775e-05,
      "loss": 1.3661,
      "mean_token_accuracy": 0.7407725945115089,
      "num_tokens": 1010621501.0,
      "step": 14060
    },
    {
      "entropy": 0.9294513553380966,
      "epoch": 3.8348323793949306,
      "grad_norm": 0.10921177268028259,
      "learning_rate": 8.887805620652594e-05,
      "loss": 1.3571,
      "mean_token_accuracy": 0.7473160177469254,
      "num_tokens": 1011342256.0,
      "step": 14070
    },
    {
      "entropy": 0.9304768949747085,
      "epoch": 3.8375579176887435,
      "grad_norm": 0.10380905121564865,
      "learning_rate": 8.885915722998397e-05,
      "loss": 1.3572,
      "mean_token_accuracy": 0.7482482895255089,
      "num_tokens": 1012074120.0,
      "step": 14080
    },
    {
      "entropy": 0.9284890264272689,
      "epoch": 3.8402834559825565,
      "grad_norm": 0.13911332190036774,
      "learning_rate": 8.884024422297554e-05,
      "loss": 1.3566,
      "mean_token_accuracy": 0.7471239194273949,
      "num_tokens": 1012793907.0,
      "step": 14090
    },
    {
      "entropy": 0.9393092960119247,
      "epoch": 3.8430089942763694,
      "grad_norm": 0.09279873222112656,
      "learning_rate": 8.882131719232931e-05,
      "loss": 1.3779,
      "mean_token_accuracy": 0.7415623143315315,
      "num_tokens": 1013501994.0,
      "step": 14100
    },
    {
      "entropy": 0.9343925014138221,
      "epoch": 3.845734532570183,
      "grad_norm": 0.09034982323646545,
      "learning_rate": 8.88023761448791e-05,
      "loss": 1.3649,
      "mean_token_accuracy": 0.7445397362112999,
      "num_tokens": 1014218259.0,
      "step": 14110
    },
    {
      "entropy": 0.9327856689691544,
      "epoch": 3.8484600708639958,
      "grad_norm": 0.0893806740641594,
      "learning_rate": 8.878342108746376e-05,
      "loss": 1.358,
      "mean_token_accuracy": 0.7467462152242661,
      "num_tokens": 1014952851.0,
      "step": 14120
    },
    {
      "entropy": 0.9390118673443795,
      "epoch": 3.8511856091578087,
      "grad_norm": 0.09976877272129059,
      "learning_rate": 8.876445202692718e-05,
      "loss": 1.3802,
      "mean_token_accuracy": 0.7416260406374932,
      "num_tokens": 1015654570.0,
      "step": 14130
    },
    {
      "entropy": 0.9356285601854324,
      "epoch": 3.8539111474516217,
      "grad_norm": 0.10243332386016846,
      "learning_rate": 8.87454689701183e-05,
      "loss": 1.375,
      "mean_token_accuracy": 0.746172621846199,
      "num_tokens": 1016358198.0,
      "step": 14140
    },
    {
      "entropy": 0.9294085711240768,
      "epoch": 3.8566366857454346,
      "grad_norm": 0.09295210987329483,
      "learning_rate": 8.872647192389116e-05,
      "loss": 1.3641,
      "mean_token_accuracy": 0.7533378571271896,
      "num_tokens": 1017094863.0,
      "step": 14150
    },
    {
      "entropy": 0.9439266920089722,
      "epoch": 3.8593622240392476,
      "grad_norm": 0.09859906136989594,
      "learning_rate": 8.870746089510483e-05,
      "loss": 1.3711,
      "mean_token_accuracy": 0.7427625700831413,
      "num_tokens": 1017813914.0,
      "step": 14160
    },
    {
      "entropy": 0.9340645805001259,
      "epoch": 3.8620877623330605,
      "grad_norm": 0.10231931507587433,
      "learning_rate": 8.868843589062339e-05,
      "loss": 1.3664,
      "mean_token_accuracy": 0.7411572426557541,
      "num_tokens": 1018514100.0,
      "step": 14170
    },
    {
      "entropy": 0.9348044991493225,
      "epoch": 3.864813300626874,
      "grad_norm": 0.14001119136810303,
      "learning_rate": 8.866939691731602e-05,
      "loss": 1.3619,
      "mean_token_accuracy": 0.7476312294602394,
      "num_tokens": 1019237095.0,
      "step": 14180
    },
    {
      "entropy": 0.9391373217105865,
      "epoch": 3.867538838920687,
      "grad_norm": 0.10235296189785004,
      "learning_rate": 8.865034398205691e-05,
      "loss": 1.3682,
      "mean_token_accuracy": 0.7420408949255943,
      "num_tokens": 1019950758.0,
      "step": 14190
    },
    {
      "entropy": 0.9349862739443779,
      "epoch": 3.8702643772145,
      "grad_norm": 0.10233399271965027,
      "learning_rate": 8.86312770917253e-05,
      "loss": 1.3654,
      "mean_token_accuracy": 0.7470587983727455,
      "num_tokens": 1020674235.0,
      "step": 14200
    },
    {
      "entropy": 0.9225364670157432,
      "epoch": 3.8729899155083127,
      "grad_norm": 0.11236976832151413,
      "learning_rate": 8.86121962532055e-05,
      "loss": 1.3612,
      "mean_token_accuracy": 0.7503849059343338,
      "num_tokens": 1021385267.0,
      "step": 14210
    },
    {
      "entropy": 0.930133692920208,
      "epoch": 3.875715453802126,
      "grad_norm": 0.10520965605974197,
      "learning_rate": 8.859310147338679e-05,
      "loss": 1.3645,
      "mean_token_accuracy": 0.7457571759819984,
      "num_tokens": 1022103810.0,
      "step": 14220
    },
    {
      "entropy": 0.9312554508447647,
      "epoch": 3.878440992095939,
      "grad_norm": 0.10042043030261993,
      "learning_rate": 8.857399275916354e-05,
      "loss": 1.3684,
      "mean_token_accuracy": 0.745972464978695,
      "num_tokens": 1022822815.0,
      "step": 14230
    },
    {
      "entropy": 0.9392946124076843,
      "epoch": 3.881166530389752,
      "grad_norm": 0.10241254419088364,
      "learning_rate": 8.855487011743514e-05,
      "loss": 1.3663,
      "mean_token_accuracy": 0.7457096457481385,
      "num_tokens": 1023545119.0,
      "step": 14240
    },
    {
      "entropy": 0.9372716471552849,
      "epoch": 3.883892068683565,
      "grad_norm": 0.09180103242397308,
      "learning_rate": 8.853573355510599e-05,
      "loss": 1.3676,
      "mean_token_accuracy": 0.7435594275593758,
      "num_tokens": 1024266085.0,
      "step": 14250
    },
    {
      "entropy": 0.9373871505260467,
      "epoch": 3.886617606977378,
      "grad_norm": 0.1257672905921936,
      "learning_rate": 8.851658307908551e-05,
      "loss": 1.3732,
      "mean_token_accuracy": 0.7408986300230026,
      "num_tokens": 1024968778.0,
      "step": 14260
    },
    {
      "entropy": 0.9347392782568932,
      "epoch": 3.889343145271191,
      "grad_norm": 0.09383546561002731,
      "learning_rate": 8.84974186962882e-05,
      "loss": 1.3648,
      "mean_token_accuracy": 0.7471752852201462,
      "num_tokens": 1025696299.0,
      "step": 14270
    },
    {
      "entropy": 0.9265581488609314,
      "epoch": 3.8920686835650042,
      "grad_norm": 0.09852515161037445,
      "learning_rate": 8.84782404136335e-05,
      "loss": 1.3576,
      "mean_token_accuracy": 0.7450048893690109,
      "num_tokens": 1026411311.0,
      "step": 14280
    },
    {
      "entropy": 0.9357007190585136,
      "epoch": 3.894794221858817,
      "grad_norm": 0.1034548357129097,
      "learning_rate": 8.845904823804595e-05,
      "loss": 1.3628,
      "mean_token_accuracy": 0.744685959815979,
      "num_tokens": 1027128276.0,
      "step": 14290
    },
    {
      "entropy": 0.9294277936220169,
      "epoch": 3.89751976015263,
      "grad_norm": 0.11570704728364944,
      "learning_rate": 8.843984217645505e-05,
      "loss": 1.3566,
      "mean_token_accuracy": 0.7450615838170052,
      "num_tokens": 1027849756.0,
      "step": 14300
    },
    {
      "entropy": 0.9366883859038353,
      "epoch": 3.900245298446443,
      "grad_norm": 0.09491495043039322,
      "learning_rate": 8.84206222357953e-05,
      "loss": 1.3731,
      "mean_token_accuracy": 0.7437934577465057,
      "num_tokens": 1028562970.0,
      "step": 14310
    },
    {
      "entropy": 0.929087245464325,
      "epoch": 3.9029708367402565,
      "grad_norm": 0.12230154126882553,
      "learning_rate": 8.840138842300628e-05,
      "loss": 1.3492,
      "mean_token_accuracy": 0.7494933113455773,
      "num_tokens": 1029286997.0,
      "step": 14320
    },
    {
      "entropy": 0.9419739976525306,
      "epoch": 3.9056963750340694,
      "grad_norm": 0.09643297642469406,
      "learning_rate": 8.838214074503254e-05,
      "loss": 1.368,
      "mean_token_accuracy": 0.7477216348052025,
      "num_tokens": 1030011594.0,
      "step": 14330
    },
    {
      "entropy": 0.9315435960888863,
      "epoch": 3.9084219133278824,
      "grad_norm": 0.1304415762424469,
      "learning_rate": 8.836287920882363e-05,
      "loss": 1.371,
      "mean_token_accuracy": 0.7494662940502167,
      "num_tokens": 1030732924.0,
      "step": 14340
    },
    {
      "entropy": 0.9234704539179802,
      "epoch": 3.9111474516216953,
      "grad_norm": 0.09431268274784088,
      "learning_rate": 8.834360382133408e-05,
      "loss": 1.3556,
      "mean_token_accuracy": 0.7517325475811958,
      "num_tokens": 1031456275.0,
      "step": 14350
    },
    {
      "entropy": 0.9329222068190575,
      "epoch": 3.9138729899155082,
      "grad_norm": 0.0933113843202591,
      "learning_rate": 8.832431458952349e-05,
      "loss": 1.3621,
      "mean_token_accuracy": 0.7461753979325294,
      "num_tokens": 1032182540.0,
      "step": 14360
    },
    {
      "entropy": 0.9314815700054169,
      "epoch": 3.916598528209321,
      "grad_norm": 0.10339117795228958,
      "learning_rate": 8.830501152035641e-05,
      "loss": 1.3678,
      "mean_token_accuracy": 0.7454869896173477,
      "num_tokens": 1032893219.0,
      "step": 14370
    },
    {
      "entropy": 0.9356389120221138,
      "epoch": 3.919324066503134,
      "grad_norm": 0.09886091947555542,
      "learning_rate": 8.828569462080238e-05,
      "loss": 1.361,
      "mean_token_accuracy": 0.7434762924909591,
      "num_tokens": 1033612556.0,
      "step": 14380
    },
    {
      "entropy": 0.928038002550602,
      "epoch": 3.9220496047969475,
      "grad_norm": 0.12218113988637924,
      "learning_rate": 8.826636389783598e-05,
      "loss": 1.3585,
      "mean_token_accuracy": 0.7478315725922584,
      "num_tokens": 1034330863.0,
      "step": 14390
    },
    {
      "entropy": 0.9243847489356994,
      "epoch": 3.9247751430907605,
      "grad_norm": 0.12371818721294403,
      "learning_rate": 8.824701935843671e-05,
      "loss": 1.3626,
      "mean_token_accuracy": 0.7491337984800339,
      "num_tokens": 1035061461.0,
      "step": 14400
    },
    {
      "entropy": 0.9274873331189155,
      "epoch": 3.9275006813845734,
      "grad_norm": 0.10390947014093399,
      "learning_rate": 8.822766100958914e-05,
      "loss": 1.363,
      "mean_token_accuracy": 0.748525820672512,
      "num_tokens": 1035778620.0,
      "step": 14410
    },
    {
      "entropy": 0.9319124907255173,
      "epoch": 3.9302262196783864,
      "grad_norm": 0.10099514573812485,
      "learning_rate": 8.820828885828276e-05,
      "loss": 1.3642,
      "mean_token_accuracy": 0.7504916310310363,
      "num_tokens": 1036503749.0,
      "step": 14420
    },
    {
      "entropy": 0.9419459164142608,
      "epoch": 3.9329517579721998,
      "grad_norm": 0.10651761293411255,
      "learning_rate": 8.818890291151207e-05,
      "loss": 1.3693,
      "mean_token_accuracy": 0.7400065645575523,
      "num_tokens": 1037207020.0,
      "step": 14430
    },
    {
      "entropy": 0.9404296949505806,
      "epoch": 3.9356772962660127,
      "grad_norm": 0.08996471017599106,
      "learning_rate": 8.816950317627654e-05,
      "loss": 1.3738,
      "mean_token_accuracy": 0.7403483390808105,
      "num_tokens": 1037918303.0,
      "step": 14440
    },
    {
      "entropy": 0.927625298500061,
      "epoch": 3.9384028345598256,
      "grad_norm": 0.09300672262907028,
      "learning_rate": 8.815008965958066e-05,
      "loss": 1.3525,
      "mean_token_accuracy": 0.750341622531414,
      "num_tokens": 1038643442.0,
      "step": 14450
    },
    {
      "entropy": 0.9373591467738152,
      "epoch": 3.9411283728536386,
      "grad_norm": 0.24935375154018402,
      "learning_rate": 8.813066236843386e-05,
      "loss": 1.3692,
      "mean_token_accuracy": 0.7430551409721374,
      "num_tokens": 1039359616.0,
      "step": 14460
    },
    {
      "entropy": 0.9297406658530235,
      "epoch": 3.9438539111474515,
      "grad_norm": 0.11255980283021927,
      "learning_rate": 8.811122130985052e-05,
      "loss": 1.3628,
      "mean_token_accuracy": 0.7447872951626777,
      "num_tokens": 1040069025.0,
      "step": 14470
    },
    {
      "entropy": 0.9282570093870163,
      "epoch": 3.9465794494412645,
      "grad_norm": 0.13645446300506592,
      "learning_rate": 8.809176649085002e-05,
      "loss": 1.366,
      "mean_token_accuracy": 0.7469138681888581,
      "num_tokens": 1040791355.0,
      "step": 14480
    },
    {
      "entropy": 0.9364361837506294,
      "epoch": 3.9493049877350774,
      "grad_norm": 0.113257497549057,
      "learning_rate": 8.807229791845673e-05,
      "loss": 1.3644,
      "mean_token_accuracy": 0.7479927495121956,
      "num_tokens": 1041520578.0,
      "step": 14490
    },
    {
      "entropy": 0.9375994071364403,
      "epoch": 3.952030526028891,
      "grad_norm": 0.09674211591482162,
      "learning_rate": 8.805281559969995e-05,
      "loss": 1.3762,
      "mean_token_accuracy": 0.7408026024699211,
      "num_tokens": 1042219593.0,
      "step": 14500
    },
    {
      "entropy": 0.933922529220581,
      "epoch": 3.9547560643227038,
      "grad_norm": 0.09420154988765717,
      "learning_rate": 8.803331954161394e-05,
      "loss": 1.367,
      "mean_token_accuracy": 0.7458790943026543,
      "num_tokens": 1042935420.0,
      "step": 14510
    },
    {
      "entropy": 0.9327138379216194,
      "epoch": 3.9574816026165167,
      "grad_norm": 0.10313500463962555,
      "learning_rate": 8.801380975123797e-05,
      "loss": 1.3548,
      "mean_token_accuracy": 0.744436514377594,
      "num_tokens": 1043655200.0,
      "step": 14520
    },
    {
      "entropy": 0.9387291252613068,
      "epoch": 3.9602071409103297,
      "grad_norm": 0.09947798401117325,
      "learning_rate": 8.799428623561621e-05,
      "loss": 1.3679,
      "mean_token_accuracy": 0.7460076659917831,
      "num_tokens": 1044382462.0,
      "step": 14530
    },
    {
      "entropy": 0.9319792449474334,
      "epoch": 3.962932679204143,
      "grad_norm": 0.1451663225889206,
      "learning_rate": 8.797474900179782e-05,
      "loss": 1.3642,
      "mean_token_accuracy": 0.7446909591555595,
      "num_tokens": 1045101795.0,
      "step": 14540
    },
    {
      "entropy": 0.942091529071331,
      "epoch": 3.965658217497956,
      "grad_norm": 0.11891796439886093,
      "learning_rate": 8.79551980568369e-05,
      "loss": 1.3747,
      "mean_token_accuracy": 0.7409000471234322,
      "num_tokens": 1045809786.0,
      "step": 14550
    },
    {
      "entropy": 0.933912581205368,
      "epoch": 3.968383755791769,
      "grad_norm": 0.11608613282442093,
      "learning_rate": 8.793563340779249e-05,
      "loss": 1.3677,
      "mean_token_accuracy": 0.745586521923542,
      "num_tokens": 1046525314.0,
      "step": 14560
    },
    {
      "entropy": 0.9301138669252396,
      "epoch": 3.971109294085582,
      "grad_norm": 0.10400774329900742,
      "learning_rate": 8.791605506172862e-05,
      "loss": 1.3655,
      "mean_token_accuracy": 0.7469741970300674,
      "num_tokens": 1047245749.0,
      "step": 14570
    },
    {
      "entropy": 0.9342464044690132,
      "epoch": 3.973834832379395,
      "grad_norm": 0.09617845714092255,
      "learning_rate": 8.789646302571422e-05,
      "loss": 1.3648,
      "mean_token_accuracy": 0.7468957230448723,
      "num_tokens": 1047970084.0,
      "step": 14580
    },
    {
      "entropy": 0.9341247498989105,
      "epoch": 3.976560370673208,
      "grad_norm": 0.0963096171617508,
      "learning_rate": 8.787685730682319e-05,
      "loss": 1.3632,
      "mean_token_accuracy": 0.745309516787529,
      "num_tokens": 1048689351.0,
      "step": 14590
    },
    {
      "entropy": 0.9294894933700562,
      "epoch": 3.9792859089670207,
      "grad_norm": 0.11984711140394211,
      "learning_rate": 8.785723791213434e-05,
      "loss": 1.3553,
      "mean_token_accuracy": 0.7462917998433113,
      "num_tokens": 1049413825.0,
      "step": 14600
    },
    {
      "entropy": 0.9345215380191803,
      "epoch": 3.982011447260834,
      "grad_norm": 0.10306016355752945,
      "learning_rate": 8.783760484873145e-05,
      "loss": 1.3705,
      "mean_token_accuracy": 0.7427061721682549,
      "num_tokens": 1050109461.0,
      "step": 14610
    },
    {
      "entropy": 0.930073706805706,
      "epoch": 3.984736985554647,
      "grad_norm": 0.10596933215856552,
      "learning_rate": 8.781795812370324e-05,
      "loss": 1.3646,
      "mean_token_accuracy": 0.749320738017559,
      "num_tokens": 1050839057.0,
      "step": 14620
    },
    {
      "entropy": 0.9409649714827537,
      "epoch": 3.98746252384846,
      "grad_norm": 0.09791397303342819,
      "learning_rate": 8.779829774414332e-05,
      "loss": 1.3768,
      "mean_token_accuracy": 0.7395751744508743,
      "num_tokens": 1051546292.0,
      "step": 14630
    },
    {
      "entropy": 0.9307864621281624,
      "epoch": 3.990188062142273,
      "grad_norm": 0.10818491876125336,
      "learning_rate": 8.777862371715026e-05,
      "loss": 1.3617,
      "mean_token_accuracy": 0.747245953977108,
      "num_tokens": 1052263190.0,
      "step": 14640
    },
    {
      "entropy": 0.9333895698189736,
      "epoch": 3.9929136004360863,
      "grad_norm": 0.0922565832734108,
      "learning_rate": 8.775893604982757e-05,
      "loss": 1.3783,
      "mean_token_accuracy": 0.7420232683420181,
      "num_tokens": 1052958401.0,
      "step": 14650
    },
    {
      "entropy": 0.9266951397061348,
      "epoch": 3.9956391387298993,
      "grad_norm": 0.09735623747110367,
      "learning_rate": 8.773923474928365e-05,
      "loss": 1.3665,
      "mean_token_accuracy": 0.7441233724355698,
      "num_tokens": 1053660786.0,
      "step": 14660
    },
    {
      "entropy": 0.9295958206057549,
      "epoch": 3.9983646770237122,
      "grad_norm": 0.10819480568170547,
      "learning_rate": 8.771951982263184e-05,
      "loss": 1.3557,
      "mean_token_accuracy": 0.7458764329552651,
      "num_tokens": 1054374956.0,
      "step": 14670
    },
    {
      "entropy": 0.9250966310501099,
      "epoch": 4.001090215317525,
      "grad_norm": 0.10195297002792358,
      "learning_rate": 8.769979127699044e-05,
      "loss": 1.3577,
      "mean_token_accuracy": 0.7487929314374924,
      "num_tokens": 1055101039.0,
      "step": 14680
    },
    {
      "entropy": 0.9206285730004311,
      "epoch": 4.003815753611338,
      "grad_norm": 0.16700603067874908,
      "learning_rate": 8.768004911948258e-05,
      "loss": 1.35,
      "mean_token_accuracy": 0.7485434412956238,
      "num_tokens": 1055816684.0,
      "step": 14690
    },
    {
      "entropy": 0.9242718994617463,
      "epoch": 4.006541291905151,
      "grad_norm": 0.10310090333223343,
      "learning_rate": 8.766029335723638e-05,
      "loss": 1.3427,
      "mean_token_accuracy": 0.7489699482917785,
      "num_tokens": 1056549268.0,
      "step": 14700
    },
    {
      "entropy": 0.9155981853604317,
      "epoch": 4.009266830198964,
      "grad_norm": 0.1116352304816246,
      "learning_rate": 8.764052399738482e-05,
      "loss": 1.3476,
      "mean_token_accuracy": 0.7441306173801422,
      "num_tokens": 1057243996.0,
      "step": 14710
    },
    {
      "entropy": 0.9187893077731133,
      "epoch": 4.011992368492777,
      "grad_norm": 0.10086751729249954,
      "learning_rate": 8.762074104706585e-05,
      "loss": 1.343,
      "mean_token_accuracy": 0.7491327404975892,
      "num_tokens": 1057971686.0,
      "step": 14720
    },
    {
      "entropy": 0.911608736217022,
      "epoch": 4.01471790678659,
      "grad_norm": 0.10362306237220764,
      "learning_rate": 8.760094451342227e-05,
      "loss": 1.345,
      "mean_token_accuracy": 0.7513741493225098,
      "num_tokens": 1058684405.0,
      "step": 14730
    },
    {
      "entropy": 0.9292451590299606,
      "epoch": 4.017443445080404,
      "grad_norm": 0.18069392442703247,
      "learning_rate": 8.75811344036018e-05,
      "loss": 1.3595,
      "mean_token_accuracy": 0.7461284607648849,
      "num_tokens": 1059403699.0,
      "step": 14740
    },
    {
      "entropy": 0.9293123945593834,
      "epoch": 4.020168983374217,
      "grad_norm": 0.10016580671072006,
      "learning_rate": 8.756131072475708e-05,
      "loss": 1.3572,
      "mean_token_accuracy": 0.7432711094617843,
      "num_tokens": 1060114983.0,
      "step": 14750
    },
    {
      "entropy": 0.9240865737199784,
      "epoch": 4.02289452166803,
      "grad_norm": 0.11918074637651443,
      "learning_rate": 8.754147348404564e-05,
      "loss": 1.3513,
      "mean_token_accuracy": 0.7447533890604973,
      "num_tokens": 1060828912.0,
      "step": 14760
    },
    {
      "entropy": 0.9221337407827377,
      "epoch": 4.025620059961843,
      "grad_norm": 0.09770968556404114,
      "learning_rate": 8.752162268862988e-05,
      "loss": 1.3521,
      "mean_token_accuracy": 0.752301150560379,
      "num_tokens": 1061564270.0,
      "step": 14770
    },
    {
      "entropy": 0.9270030155777931,
      "epoch": 4.0283455982556555,
      "grad_norm": 0.10164887458086014,
      "learning_rate": 8.750175834567715e-05,
      "loss": 1.3531,
      "mean_token_accuracy": 0.7463859170675278,
      "num_tokens": 1062281406.0,
      "step": 14780
    },
    {
      "entropy": 0.928347361087799,
      "epoch": 4.0310711365494685,
      "grad_norm": 0.10720933973789215,
      "learning_rate": 8.748188046235962e-05,
      "loss": 1.3576,
      "mean_token_accuracy": 0.7476789802312851,
      "num_tokens": 1062988575.0,
      "step": 14790
    },
    {
      "entropy": 0.9258315578103066,
      "epoch": 4.033796674843281,
      "grad_norm": 0.10308457911014557,
      "learning_rate": 8.746198904585443e-05,
      "loss": 1.3606,
      "mean_token_accuracy": 0.7435396477580071,
      "num_tokens": 1063688290.0,
      "step": 14800
    },
    {
      "entropy": 0.9214114010334015,
      "epoch": 4.036522213137094,
      "grad_norm": 0.09378824383020401,
      "learning_rate": 8.744208410334353e-05,
      "loss": 1.3483,
      "mean_token_accuracy": 0.7454946652054787,
      "num_tokens": 1064391818.0,
      "step": 14810
    },
    {
      "entropy": 0.9201286911964417,
      "epoch": 4.039247751430907,
      "grad_norm": 0.1108425185084343,
      "learning_rate": 8.742216564201382e-05,
      "loss": 1.3423,
      "mean_token_accuracy": 0.7523040413856507,
      "num_tokens": 1065132210.0,
      "step": 14820
    },
    {
      "entropy": 0.9173980921506881,
      "epoch": 4.04197328972472,
      "grad_norm": 0.10384967923164368,
      "learning_rate": 8.740223366905702e-05,
      "loss": 1.3471,
      "mean_token_accuracy": 0.7506538406014442,
      "num_tokens": 1065850549.0,
      "step": 14830
    },
    {
      "entropy": 0.9249087899923325,
      "epoch": 4.044698828018534,
      "grad_norm": 0.10688713192939758,
      "learning_rate": 8.738228819166979e-05,
      "loss": 1.3553,
      "mean_token_accuracy": 0.7454244300723076,
      "num_tokens": 1066557825.0,
      "step": 14840
    },
    {
      "entropy": 0.9281580656766891,
      "epoch": 4.047424366312347,
      "grad_norm": 0.09900780022144318,
      "learning_rate": 8.736232921705358e-05,
      "loss": 1.3537,
      "mean_token_accuracy": 0.7471346974372863,
      "num_tokens": 1067280883.0,
      "step": 14850
    },
    {
      "entropy": 0.9204226359724998,
      "epoch": 4.05014990460616,
      "grad_norm": 0.1197301372885704,
      "learning_rate": 8.734235675241482e-05,
      "loss": 1.344,
      "mean_token_accuracy": 0.7487455755472183,
      "num_tokens": 1068005115.0,
      "step": 14860
    },
    {
      "entropy": 0.9192662566900254,
      "epoch": 4.052875442899973,
      "grad_norm": 0.10520618408918381,
      "learning_rate": 8.732237080496474e-05,
      "loss": 1.3424,
      "mean_token_accuracy": 0.7479520559310913,
      "num_tokens": 1068729708.0,
      "step": 14870
    },
    {
      "entropy": 0.9146839618682862,
      "epoch": 4.055600981193786,
      "grad_norm": 0.09667209535837173,
      "learning_rate": 8.730237138191944e-05,
      "loss": 1.3497,
      "mean_token_accuracy": 0.7494596183300019,
      "num_tokens": 1069445170.0,
      "step": 14880
    },
    {
      "entropy": 0.922248686850071,
      "epoch": 4.058326519487599,
      "grad_norm": 0.10695157200098038,
      "learning_rate": 8.728235849049992e-05,
      "loss": 1.358,
      "mean_token_accuracy": 0.7500324621796608,
      "num_tokens": 1070172712.0,
      "step": 14890
    },
    {
      "entropy": 0.9240766763687134,
      "epoch": 4.061052057781412,
      "grad_norm": 0.09809698164463043,
      "learning_rate": 8.7262332137932e-05,
      "loss": 1.3626,
      "mean_token_accuracy": 0.7443317130208016,
      "num_tokens": 1070872748.0,
      "step": 14900
    },
    {
      "entropy": 0.9219892278313637,
      "epoch": 4.063777596075225,
      "grad_norm": 0.10685938596725464,
      "learning_rate": 8.724229233144643e-05,
      "loss": 1.3541,
      "mean_token_accuracy": 0.7486927181482315,
      "num_tokens": 1071581000.0,
      "step": 14910
    },
    {
      "entropy": 0.925316508114338,
      "epoch": 4.066503134369038,
      "grad_norm": 0.11054862290620804,
      "learning_rate": 8.722223907827873e-05,
      "loss": 1.3561,
      "mean_token_accuracy": 0.7469108149409294,
      "num_tokens": 1072306023.0,
      "step": 14920
    },
    {
      "entropy": 0.9275324031710624,
      "epoch": 4.069228672662851,
      "grad_norm": 0.09480039775371552,
      "learning_rate": 8.720217238566932e-05,
      "loss": 1.361,
      "mean_token_accuracy": 0.745964989066124,
      "num_tokens": 1073018851.0,
      "step": 14930
    },
    {
      "entropy": 0.9243619382381439,
      "epoch": 4.0719542109566635,
      "grad_norm": 0.10341368615627289,
      "learning_rate": 8.718209226086349e-05,
      "loss": 1.3491,
      "mean_token_accuracy": 0.7486465230584145,
      "num_tokens": 1073730459.0,
      "step": 14940
    },
    {
      "entropy": 0.923155564069748,
      "epoch": 4.074679749250477,
      "grad_norm": 0.09840913861989975,
      "learning_rate": 8.716199871111134e-05,
      "loss": 1.3464,
      "mean_token_accuracy": 0.7452421724796295,
      "num_tokens": 1074448014.0,
      "step": 14950
    },
    {
      "entropy": 0.9233433306217194,
      "epoch": 4.07740528754429,
      "grad_norm": 0.09292308986186981,
      "learning_rate": 8.714189174366785e-05,
      "loss": 1.3558,
      "mean_token_accuracy": 0.7448950618505478,
      "num_tokens": 1075147180.0,
      "step": 14960
    },
    {
      "entropy": 0.9276746451854706,
      "epoch": 4.080130825838103,
      "grad_norm": 0.09625004976987839,
      "learning_rate": 8.712177136579282e-05,
      "loss": 1.353,
      "mean_token_accuracy": 0.7442156180739403,
      "num_tokens": 1075857920.0,
      "step": 14970
    },
    {
      "entropy": 0.9259742245078086,
      "epoch": 4.082856364131916,
      "grad_norm": 0.09763691574335098,
      "learning_rate": 8.710163758475091e-05,
      "loss": 1.3558,
      "mean_token_accuracy": 0.7508269518613815,
      "num_tokens": 1076585194.0,
      "step": 14980
    },
    {
      "entropy": 0.9270025014877319,
      "epoch": 4.085581902425729,
      "grad_norm": 0.09863530844449997,
      "learning_rate": 8.708149040781161e-05,
      "loss": 1.3524,
      "mean_token_accuracy": 0.7452347114682197,
      "num_tokens": 1077297977.0,
      "step": 14990
    },
    {
      "entropy": 0.9185169100761413,
      "epoch": 4.088307440719542,
      "grad_norm": 0.11759143322706223,
      "learning_rate": 8.706132984224926e-05,
      "loss": 1.3544,
      "mean_token_accuracy": 0.7505017936229705,
      "num_tokens": 1078016764.0,
      "step": 15000
    },
    {
      "entropy": 0.9283453524112701,
      "epoch": 4.091032979013355,
      "grad_norm": 0.11131831258535385,
      "learning_rate": 8.7041155895343e-05,
      "loss": 1.3527,
      "mean_token_accuracy": 0.7450633034110069,
      "num_tokens": 1078731211.0,
      "step": 15010
    },
    {
      "entropy": 0.9266245886683464,
      "epoch": 4.093758517307168,
      "grad_norm": 0.10359904170036316,
      "learning_rate": 8.702096857437685e-05,
      "loss": 1.3593,
      "mean_token_accuracy": 0.7451131463050842,
      "num_tokens": 1079442047.0,
      "step": 15020
    },
    {
      "entropy": 0.9161438152194024,
      "epoch": 4.096484055600981,
      "grad_norm": 0.1034112200140953,
      "learning_rate": 8.700076788663961e-05,
      "loss": 1.3477,
      "mean_token_accuracy": 0.7492471635341644,
      "num_tokens": 1080154598.0,
      "step": 15030
    },
    {
      "entropy": 0.9216574400663375,
      "epoch": 4.099209593894794,
      "grad_norm": 0.11114446073770523,
      "learning_rate": 8.698055383942493e-05,
      "loss": 1.3542,
      "mean_token_accuracy": 0.7474159732460975,
      "num_tokens": 1080867093.0,
      "step": 15040
    },
    {
      "entropy": 0.923461177945137,
      "epoch": 4.101935132188607,
      "grad_norm": 0.1069541946053505,
      "learning_rate": 8.696032644003132e-05,
      "loss": 1.3545,
      "mean_token_accuracy": 0.7468228310346603,
      "num_tokens": 1081582715.0,
      "step": 15050
    },
    {
      "entropy": 0.920073664188385,
      "epoch": 4.104660670482421,
      "grad_norm": 0.2267809361219406,
      "learning_rate": 8.694008569576201e-05,
      "loss": 1.3522,
      "mean_token_accuracy": 0.7503765150904655,
      "num_tokens": 1082308456.0,
      "step": 15060
    },
    {
      "entropy": 0.9166083291172982,
      "epoch": 4.107386208776234,
      "grad_norm": 0.10917073488235474,
      "learning_rate": 8.691983161392518e-05,
      "loss": 1.3392,
      "mean_token_accuracy": 0.7507751628756523,
      "num_tokens": 1083039344.0,
      "step": 15070
    },
    {
      "entropy": 0.9270202234387398,
      "epoch": 4.110111747070047,
      "grad_norm": 0.10152757167816162,
      "learning_rate": 8.689956420183371e-05,
      "loss": 1.3555,
      "mean_token_accuracy": 0.7467972695827484,
      "num_tokens": 1083757478.0,
      "step": 15080
    },
    {
      "entropy": 0.919314056634903,
      "epoch": 4.1128372853638595,
      "grad_norm": 0.12157361954450607,
      "learning_rate": 8.687928346680537e-05,
      "loss": 1.3462,
      "mean_token_accuracy": 0.7490699544548989,
      "num_tokens": 1084476929.0,
      "step": 15090
    },
    {
      "entropy": 0.9161649614572525,
      "epoch": 4.1155628236576725,
      "grad_norm": 0.10362573713064194,
      "learning_rate": 8.685898941616269e-05,
      "loss": 1.3546,
      "mean_token_accuracy": 0.7478635162115097,
      "num_tokens": 1085189812.0,
      "step": 15100
    },
    {
      "entropy": 0.9304158881306648,
      "epoch": 4.118288361951485,
      "grad_norm": 0.10824136435985565,
      "learning_rate": 8.683868205723305e-05,
      "loss": 1.349,
      "mean_token_accuracy": 0.7461272209882737,
      "num_tokens": 1085916495.0,
      "step": 15110
    },
    {
      "entropy": 0.9167304843664169,
      "epoch": 4.121013900245298,
      "grad_norm": 0.10715682804584503,
      "learning_rate": 8.681836139734858e-05,
      "loss": 1.3414,
      "mean_token_accuracy": 0.7514822587370873,
      "num_tokens": 1086636334.0,
      "step": 15120
    },
    {
      "entropy": 0.9281087353825569,
      "epoch": 4.123739438539111,
      "grad_norm": 0.09745436906814575,
      "learning_rate": 8.679802744384629e-05,
      "loss": 1.3502,
      "mean_token_accuracy": 0.7524097695946693,
      "num_tokens": 1087383018.0,
      "step": 15130
    },
    {
      "entropy": 0.9187511190772056,
      "epoch": 4.126464976832924,
      "grad_norm": 0.12303081154823303,
      "learning_rate": 8.677768020406792e-05,
      "loss": 1.3427,
      "mean_token_accuracy": 0.7529912620782853,
      "num_tokens": 1088117232.0,
      "step": 15140
    },
    {
      "entropy": 0.9229072242975235,
      "epoch": 4.129190515126737,
      "grad_norm": 0.10220523923635483,
      "learning_rate": 8.675731968536002e-05,
      "loss": 1.3603,
      "mean_token_accuracy": 0.7439503133296966,
      "num_tokens": 1088819662.0,
      "step": 15150
    },
    {
      "entropy": 0.9190058186650276,
      "epoch": 4.13191605342055,
      "grad_norm": 0.10615793615579605,
      "learning_rate": 8.673694589507398e-05,
      "loss": 1.3538,
      "mean_token_accuracy": 0.7525135144591332,
      "num_tokens": 1089546727.0,
      "step": 15160
    },
    {
      "entropy": 0.9210917025804519,
      "epoch": 4.134641591714364,
      "grad_norm": 0.09824615716934204,
      "learning_rate": 8.671655884056595e-05,
      "loss": 1.3547,
      "mean_token_accuracy": 0.7447015568614006,
      "num_tokens": 1090256419.0,
      "step": 15170
    },
    {
      "entropy": 0.9234574690461159,
      "epoch": 4.137367130008177,
      "grad_norm": 0.09964392334222794,
      "learning_rate": 8.669615852919683e-05,
      "loss": 1.3533,
      "mean_token_accuracy": 0.750318144261837,
      "num_tokens": 1090980564.0,
      "step": 15180
    },
    {
      "entropy": 0.9199197620153428,
      "epoch": 4.14009266830199,
      "grad_norm": 0.10636098682880402,
      "learning_rate": 8.667574496833239e-05,
      "loss": 1.345,
      "mean_token_accuracy": 0.7478230789303779,
      "num_tokens": 1091698581.0,
      "step": 15190
    },
    {
      "entropy": 0.9234969943761826,
      "epoch": 4.142818206595803,
      "grad_norm": 0.09744471311569214,
      "learning_rate": 8.665531816534311e-05,
      "loss": 1.3663,
      "mean_token_accuracy": 0.7447913825511933,
      "num_tokens": 1092398547.0,
      "step": 15200
    },
    {
      "entropy": 0.9177030801773072,
      "epoch": 4.145543744889616,
      "grad_norm": 0.1044195294380188,
      "learning_rate": 8.66348781276043e-05,
      "loss": 1.3466,
      "mean_token_accuracy": 0.7534613877534866,
      "num_tokens": 1093129474.0,
      "step": 15210
    },
    {
      "entropy": 0.925013567507267,
      "epoch": 4.148269283183429,
      "grad_norm": 0.11116165667772293,
      "learning_rate": 8.661442486249602e-05,
      "loss": 1.3572,
      "mean_token_accuracy": 0.7482627645134926,
      "num_tokens": 1093854917.0,
      "step": 15220
    },
    {
      "entropy": 0.9209140434861183,
      "epoch": 4.150994821477242,
      "grad_norm": 0.10993218421936035,
      "learning_rate": 8.65939583774031e-05,
      "loss": 1.3534,
      "mean_token_accuracy": 0.7482733398675918,
      "num_tokens": 1094585239.0,
      "step": 15230
    },
    {
      "entropy": 0.9196558341383934,
      "epoch": 4.153720359771055,
      "grad_norm": 0.10569615662097931,
      "learning_rate": 8.65734786797152e-05,
      "loss": 1.3531,
      "mean_token_accuracy": 0.7470766916871071,
      "num_tokens": 1095288938.0,
      "step": 15240
    },
    {
      "entropy": 0.9233587205410003,
      "epoch": 4.1564458980648675,
      "grad_norm": 0.10359329730272293,
      "learning_rate": 8.655298577682664e-05,
      "loss": 1.3525,
      "mean_token_accuracy": 0.7450036704540253,
      "num_tokens": 1096006276.0,
      "step": 15250
    },
    {
      "entropy": 0.9250382363796235,
      "epoch": 4.1591714363586805,
      "grad_norm": 0.10726846009492874,
      "learning_rate": 8.653247967613665e-05,
      "loss": 1.3464,
      "mean_token_accuracy": 0.7515286386013031,
      "num_tokens": 1096741770.0,
      "step": 15260
    },
    {
      "entropy": 0.9240881979465485,
      "epoch": 4.161896974652494,
      "grad_norm": 0.10148107260465622,
      "learning_rate": 8.651196038504911e-05,
      "loss": 1.3516,
      "mean_token_accuracy": 0.7479907020926475,
      "num_tokens": 1097462882.0,
      "step": 15270
    },
    {
      "entropy": 0.9232209980487823,
      "epoch": 4.164622512946307,
      "grad_norm": 0.10852691531181335,
      "learning_rate": 8.649142791097272e-05,
      "loss": 1.3573,
      "mean_token_accuracy": 0.7463250428438186,
      "num_tokens": 1098171286.0,
      "step": 15280
    },
    {
      "entropy": 0.9185991257429122,
      "epoch": 4.16734805124012,
      "grad_norm": 0.10770787298679352,
      "learning_rate": 8.647088226132092e-05,
      "loss": 1.3545,
      "mean_token_accuracy": 0.7466697439551353,
      "num_tokens": 1098874143.0,
      "step": 15290
    },
    {
      "entropy": 0.9250710621476174,
      "epoch": 4.170073589533933,
      "grad_norm": 0.10845988243818283,
      "learning_rate": 8.645032344351191e-05,
      "loss": 1.3602,
      "mean_token_accuracy": 0.7440716817975044,
      "num_tokens": 1099575728.0,
      "step": 15300
    },
    {
      "entropy": 0.9233541503548622,
      "epoch": 4.172799127827746,
      "grad_norm": 0.10754302144050598,
      "learning_rate": 8.642975146496863e-05,
      "loss": 1.349,
      "mean_token_accuracy": 0.746393920481205,
      "num_tokens": 1100286300.0,
      "step": 15310
    },
    {
      "entropy": 0.925643177330494,
      "epoch": 4.175524666121559,
      "grad_norm": 0.11739949882030487,
      "learning_rate": 8.640916633311882e-05,
      "loss": 1.3605,
      "mean_token_accuracy": 0.7466436505317688,
      "num_tokens": 1101007355.0,
      "step": 15320
    },
    {
      "entropy": 0.9255161434412003,
      "epoch": 4.178250204415372,
      "grad_norm": 0.10572593659162521,
      "learning_rate": 8.638856805539492e-05,
      "loss": 1.3573,
      "mean_token_accuracy": 0.7447706922888756,
      "num_tokens": 1101712303.0,
      "step": 15330
    },
    {
      "entropy": 0.9263571187853813,
      "epoch": 4.180975742709185,
      "grad_norm": 0.11344701051712036,
      "learning_rate": 8.636795663923412e-05,
      "loss": 1.3564,
      "mean_token_accuracy": 0.7460677474737167,
      "num_tokens": 1102429425.0,
      "step": 15340
    },
    {
      "entropy": 0.9213182166218757,
      "epoch": 4.183701281002998,
      "grad_norm": 0.09633529931306839,
      "learning_rate": 8.634733209207838e-05,
      "loss": 1.3533,
      "mean_token_accuracy": 0.7468613177537918,
      "num_tokens": 1103142940.0,
      "step": 15350
    },
    {
      "entropy": 0.9199415788054466,
      "epoch": 4.186426819296811,
      "grad_norm": 0.08518577367067337,
      "learning_rate": 8.63266944213744e-05,
      "loss": 1.3389,
      "mean_token_accuracy": 0.7497214883565902,
      "num_tokens": 1103873112.0,
      "step": 15360
    },
    {
      "entropy": 0.9211140155792237,
      "epoch": 4.189152357590624,
      "grad_norm": 0.09612301737070084,
      "learning_rate": 8.630604363457357e-05,
      "loss": 1.348,
      "mean_token_accuracy": 0.7492274045944214,
      "num_tokens": 1104595926.0,
      "step": 15370
    },
    {
      "entropy": 0.9184622243046761,
      "epoch": 4.191877895884438,
      "grad_norm": 0.10593047738075256,
      "learning_rate": 8.62853797391321e-05,
      "loss": 1.3515,
      "mean_token_accuracy": 0.7543678924441337,
      "num_tokens": 1105332142.0,
      "step": 15380
    },
    {
      "entropy": 0.919635309278965,
      "epoch": 4.1946034341782505,
      "grad_norm": 0.09958937019109726,
      "learning_rate": 8.626470274251086e-05,
      "loss": 1.3517,
      "mean_token_accuracy": 0.7483934178948403,
      "num_tokens": 1106054683.0,
      "step": 15390
    },
    {
      "entropy": 0.9255688995122909,
      "epoch": 4.1973289724720635,
      "grad_norm": 0.09910338371992111,
      "learning_rate": 8.624401265217546e-05,
      "loss": 1.3539,
      "mean_token_accuracy": 0.7464926168322563,
      "num_tokens": 1106775141.0,
      "step": 15400
    },
    {
      "entropy": 0.9133437097072601,
      "epoch": 4.200054510765876,
      "grad_norm": 0.10686346143484116,
      "learning_rate": 8.622330947559626e-05,
      "loss": 1.3387,
      "mean_token_accuracy": 0.7509348660707473,
      "num_tokens": 1107500783.0,
      "step": 15410
    },
    {
      "entropy": 0.9259075611829758,
      "epoch": 4.202780049059689,
      "grad_norm": 0.10239091515541077,
      "learning_rate": 8.620259322024837e-05,
      "loss": 1.3568,
      "mean_token_accuracy": 0.7503230765461921,
      "num_tokens": 1108223072.0,
      "step": 15420
    },
    {
      "entropy": 0.9259609490633011,
      "epoch": 4.205505587353502,
      "grad_norm": 0.09917110204696655,
      "learning_rate": 8.618186389361157e-05,
      "loss": 1.349,
      "mean_token_accuracy": 0.7449682936072349,
      "num_tokens": 1108939663.0,
      "step": 15430
    },
    {
      "entropy": 0.9250614657998085,
      "epoch": 4.208231125647315,
      "grad_norm": 0.10166233777999878,
      "learning_rate": 8.616112150317038e-05,
      "loss": 1.3583,
      "mean_token_accuracy": 0.7445510044693947,
      "num_tokens": 1109643749.0,
      "step": 15440
    },
    {
      "entropy": 0.9110371172428131,
      "epoch": 4.210956663941128,
      "grad_norm": 0.10801061242818832,
      "learning_rate": 8.614036605641401e-05,
      "loss": 1.3493,
      "mean_token_accuracy": 0.7497808769345283,
      "num_tokens": 1110353578.0,
      "step": 15450
    },
    {
      "entropy": 0.9236445412039757,
      "epoch": 4.213682202234941,
      "grad_norm": 0.106284499168396,
      "learning_rate": 8.611959756083645e-05,
      "loss": 1.3523,
      "mean_token_accuracy": 0.7495973750948906,
      "num_tokens": 1111078312.0,
      "step": 15460
    },
    {
      "entropy": 0.9197873383760452,
      "epoch": 4.216407740528754,
      "grad_norm": 0.1084304079413414,
      "learning_rate": 8.609881602393632e-05,
      "loss": 1.3536,
      "mean_token_accuracy": 0.7510551482439041,
      "num_tokens": 1111800138.0,
      "step": 15470
    },
    {
      "entropy": 0.9248720005154609,
      "epoch": 4.219133278822567,
      "grad_norm": 0.09371249377727509,
      "learning_rate": 8.607802145321704e-05,
      "loss": 1.3568,
      "mean_token_accuracy": 0.7465752214193344,
      "num_tokens": 1112506023.0,
      "step": 15480
    },
    {
      "entropy": 0.918645641207695,
      "epoch": 4.221858817116381,
      "grad_norm": 0.10070053488016129,
      "learning_rate": 8.605721385618666e-05,
      "loss": 1.3527,
      "mean_token_accuracy": 0.7517626136541367,
      "num_tokens": 1113229056.0,
      "step": 15490
    },
    {
      "entropy": 0.9234529182314872,
      "epoch": 4.224584355410194,
      "grad_norm": 0.10140307247638702,
      "learning_rate": 8.603639324035796e-05,
      "loss": 1.3529,
      "mean_token_accuracy": 0.7472596898674965,
      "num_tokens": 1113939386.0,
      "step": 15500
    },
    {
      "entropy": 0.9218262240290642,
      "epoch": 4.227309893704007,
      "grad_norm": 0.11644724756479263,
      "learning_rate": 8.601555961324844e-05,
      "loss": 1.3496,
      "mean_token_accuracy": 0.7462103083729744,
      "num_tokens": 1114665936.0,
      "step": 15510
    },
    {
      "entropy": 0.9324785009026527,
      "epoch": 4.23003543199782,
      "grad_norm": 0.11617831885814667,
      "learning_rate": 8.599471298238022e-05,
      "loss": 1.3581,
      "mean_token_accuracy": 0.743926165997982,
      "num_tokens": 1115377509.0,
      "step": 15520
    },
    {
      "entropy": 0.9201160460710526,
      "epoch": 4.232760970291633,
      "grad_norm": 0.11622854322195053,
      "learning_rate": 8.597385335528024e-05,
      "loss": 1.3401,
      "mean_token_accuracy": 0.7489723354578018,
      "num_tokens": 1116105358.0,
      "step": 15530
    },
    {
      "entropy": 0.9292018488049507,
      "epoch": 4.235486508585446,
      "grad_norm": 0.10608626902103424,
      "learning_rate": 8.595298073948003e-05,
      "loss": 1.3552,
      "mean_token_accuracy": 0.7496816053986549,
      "num_tokens": 1116836485.0,
      "step": 15540
    },
    {
      "entropy": 0.9256255522370338,
      "epoch": 4.238212046879259,
      "grad_norm": 0.11927047371864319,
      "learning_rate": 8.593209514251587e-05,
      "loss": 1.3568,
      "mean_token_accuracy": 0.7435988053679466,
      "num_tokens": 1117542989.0,
      "step": 15550
    },
    {
      "entropy": 0.9222154036164284,
      "epoch": 4.2409375851730715,
      "grad_norm": 0.11140109598636627,
      "learning_rate": 8.591119657192868e-05,
      "loss": 1.3539,
      "mean_token_accuracy": 0.7469932600855828,
      "num_tokens": 1118253384.0,
      "step": 15560
    },
    {
      "entropy": 0.921469333767891,
      "epoch": 4.2436631234668845,
      "grad_norm": 0.10192237049341202,
      "learning_rate": 8.58902850352641e-05,
      "loss": 1.3462,
      "mean_token_accuracy": 0.7476293072104454,
      "num_tokens": 1118976778.0,
      "step": 15570
    },
    {
      "entropy": 0.9227729007601738,
      "epoch": 4.246388661760697,
      "grad_norm": 0.094509556889534,
      "learning_rate": 8.586936054007243e-05,
      "loss": 1.3542,
      "mean_token_accuracy": 0.7453227132558823,
      "num_tokens": 1119690904.0,
      "step": 15580
    },
    {
      "entropy": 0.9180780068039894,
      "epoch": 4.24911420005451,
      "grad_norm": 0.09690500050783157,
      "learning_rate": 8.584842309390866e-05,
      "loss": 1.3337,
      "mean_token_accuracy": 0.7538761734962464,
      "num_tokens": 1120416660.0,
      "step": 15590
    },
    {
      "entropy": 0.9172228232026101,
      "epoch": 4.251839738348324,
      "grad_norm": 0.096047043800354,
      "learning_rate": 8.582747270433244e-05,
      "loss": 1.3481,
      "mean_token_accuracy": 0.748733003437519,
      "num_tokens": 1121124690.0,
      "step": 15600
    },
    {
      "entropy": 0.917376683652401,
      "epoch": 4.254565276642137,
      "grad_norm": 0.09743868559598923,
      "learning_rate": 8.580650937890814e-05,
      "loss": 1.3523,
      "mean_token_accuracy": 0.7458606168627739,
      "num_tokens": 1121840947.0,
      "step": 15610
    },
    {
      "entropy": 0.9286151498556137,
      "epoch": 4.25729081493595,
      "grad_norm": 0.10876942425966263,
      "learning_rate": 8.578553312520473e-05,
      "loss": 1.3642,
      "mean_token_accuracy": 0.7437870189547539,
      "num_tokens": 1122550011.0,
      "step": 15620
    },
    {
      "entropy": 0.9214182883501053,
      "epoch": 4.260016353229763,
      "grad_norm": 0.13369660079479218,
      "learning_rate": 8.576454395079592e-05,
      "loss": 1.3448,
      "mean_token_accuracy": 0.748539000749588,
      "num_tokens": 1123263873.0,
      "step": 15630
    },
    {
      "entropy": 0.9182384103536606,
      "epoch": 4.262741891523576,
      "grad_norm": 0.10016344487667084,
      "learning_rate": 8.574354186326001e-05,
      "loss": 1.3527,
      "mean_token_accuracy": 0.7461849331855774,
      "num_tokens": 1123977611.0,
      "step": 15640
    },
    {
      "entropy": 0.9246586412191391,
      "epoch": 4.265467429817389,
      "grad_norm": 0.09606947004795074,
      "learning_rate": 8.572252687018003e-05,
      "loss": 1.3618,
      "mean_token_accuracy": 0.7431387960910797,
      "num_tokens": 1124676896.0,
      "step": 15650
    },
    {
      "entropy": 0.9174091562628746,
      "epoch": 4.268192968111202,
      "grad_norm": 0.11599027365446091,
      "learning_rate": 8.570149897914365e-05,
      "loss": 1.3441,
      "mean_token_accuracy": 0.7503603219985961,
      "num_tokens": 1125394167.0,
      "step": 15660
    },
    {
      "entropy": 0.9210040450096131,
      "epoch": 4.270918506405015,
      "grad_norm": 0.12634681165218353,
      "learning_rate": 8.568045819774317e-05,
      "loss": 1.3536,
      "mean_token_accuracy": 0.750414302945137,
      "num_tokens": 1126120943.0,
      "step": 15670
    },
    {
      "entropy": 0.9279288172721862,
      "epoch": 4.273644044698828,
      "grad_norm": 0.11021529138088226,
      "learning_rate": 8.565940453357556e-05,
      "loss": 1.3602,
      "mean_token_accuracy": 0.7447734594345092,
      "num_tokens": 1126837097.0,
      "step": 15680
    },
    {
      "entropy": 0.925048041343689,
      "epoch": 4.276369582992641,
      "grad_norm": 0.10797903686761856,
      "learning_rate": 8.563833799424245e-05,
      "loss": 1.3567,
      "mean_token_accuracy": 0.7467201858758926,
      "num_tokens": 1127556011.0,
      "step": 15690
    },
    {
      "entropy": 0.9306846410036087,
      "epoch": 4.2790951212864545,
      "grad_norm": 0.10243770480155945,
      "learning_rate": 8.561725858735012e-05,
      "loss": 1.353,
      "mean_token_accuracy": 0.7427692621946335,
      "num_tokens": 1128276613.0,
      "step": 15700
    },
    {
      "entropy": 0.919539888203144,
      "epoch": 4.2818206595802675,
      "grad_norm": 0.10891085118055344,
      "learning_rate": 8.55961663205095e-05,
      "loss": 1.3594,
      "mean_token_accuracy": 0.743375138938427,
      "num_tokens": 1128968352.0,
      "step": 15710
    },
    {
      "entropy": 0.9251904681324958,
      "epoch": 4.28454619787408,
      "grad_norm": 0.10192779451608658,
      "learning_rate": 8.557506120133612e-05,
      "loss": 1.3535,
      "mean_token_accuracy": 0.7490128025412559,
      "num_tokens": 1129695155.0,
      "step": 15720
    },
    {
      "entropy": 0.9143778622150421,
      "epoch": 4.287271736167893,
      "grad_norm": 0.10731621831655502,
      "learning_rate": 8.555394323745021e-05,
      "loss": 1.3479,
      "mean_token_accuracy": 0.7502439513802528,
      "num_tokens": 1130414342.0,
      "step": 15730
    },
    {
      "entropy": 0.9239508718252182,
      "epoch": 4.289997274461706,
      "grad_norm": 0.10015898197889328,
      "learning_rate": 8.553281243647661e-05,
      "loss": 1.3522,
      "mean_token_accuracy": 0.7509477585554123,
      "num_tokens": 1131144199.0,
      "step": 15740
    },
    {
      "entropy": 0.9253065183758735,
      "epoch": 4.292722812755519,
      "grad_norm": 0.10033026337623596,
      "learning_rate": 8.551166880604478e-05,
      "loss": 1.3535,
      "mean_token_accuracy": 0.7482062309980393,
      "num_tokens": 1131872682.0,
      "step": 15750
    },
    {
      "entropy": 0.933468659222126,
      "epoch": 4.295448351049332,
      "grad_norm": 0.10383269190788269,
      "learning_rate": 8.549051235378883e-05,
      "loss": 1.3573,
      "mean_token_accuracy": 0.7434056118130684,
      "num_tokens": 1132586135.0,
      "step": 15760
    },
    {
      "entropy": 0.9203952506184578,
      "epoch": 4.298173889343145,
      "grad_norm": 0.10781288146972656,
      "learning_rate": 8.54693430873475e-05,
      "loss": 1.3545,
      "mean_token_accuracy": 0.752930548787117,
      "num_tokens": 1133323703.0,
      "step": 15770
    },
    {
      "entropy": 0.9198941394686699,
      "epoch": 4.300899427636958,
      "grad_norm": 0.09412723779678345,
      "learning_rate": 8.544816101436417e-05,
      "loss": 1.3517,
      "mean_token_accuracy": 0.7454964980483055,
      "num_tokens": 1134030538.0,
      "step": 15780
    },
    {
      "entropy": 0.9307396382093429,
      "epoch": 4.303624965930771,
      "grad_norm": 0.09997003525495529,
      "learning_rate": 8.542696614248682e-05,
      "loss": 1.3584,
      "mean_token_accuracy": 0.7488197386264801,
      "num_tokens": 1134764208.0,
      "step": 15790
    },
    {
      "entropy": 0.9198611840605736,
      "epoch": 4.306350504224584,
      "grad_norm": 0.12314765155315399,
      "learning_rate": 8.540575847936803e-05,
      "loss": 1.356,
      "mean_token_accuracy": 0.7507635086774826,
      "num_tokens": 1135481334.0,
      "step": 15800
    },
    {
      "entropy": 0.9221617951989174,
      "epoch": 4.309076042518397,
      "grad_norm": 0.10982570052146912,
      "learning_rate": 8.538453803266507e-05,
      "loss": 1.3481,
      "mean_token_accuracy": 0.7472605004906654,
      "num_tokens": 1136204421.0,
      "step": 15810
    },
    {
      "entropy": 0.9230554610490799,
      "epoch": 4.311801580812211,
      "grad_norm": 0.1046750396490097,
      "learning_rate": 8.536330481003976e-05,
      "loss": 1.3513,
      "mean_token_accuracy": 0.7465279966592788,
      "num_tokens": 1136924603.0,
      "step": 15820
    },
    {
      "entropy": 0.9156743109226226,
      "epoch": 4.314527119106024,
      "grad_norm": 0.10535377264022827,
      "learning_rate": 8.534205881915857e-05,
      "loss": 1.3465,
      "mean_token_accuracy": 0.7513804316520691,
      "num_tokens": 1137650525.0,
      "step": 15830
    },
    {
      "entropy": 0.9267956629395485,
      "epoch": 4.317252657399837,
      "grad_norm": 0.11267965286970139,
      "learning_rate": 8.532080006769255e-05,
      "loss": 1.3664,
      "mean_token_accuracy": 0.7408849030733109,
      "num_tokens": 1138347900.0,
      "step": 15840
    },
    {
      "entropy": 0.92094816416502,
      "epoch": 4.31997819569365,
      "grad_norm": 0.10393417626619339,
      "learning_rate": 8.529952856331738e-05,
      "loss": 1.3562,
      "mean_token_accuracy": 0.7493443578481674,
      "num_tokens": 1139067471.0,
      "step": 15850
    },
    {
      "entropy": 0.9262821987271309,
      "epoch": 4.3227037339874625,
      "grad_norm": 0.10966341197490692,
      "learning_rate": 8.527824431371332e-05,
      "loss": 1.3571,
      "mean_token_accuracy": 0.746008823812008,
      "num_tokens": 1139784535.0,
      "step": 15860
    },
    {
      "entropy": 0.9217200368642807,
      "epoch": 4.3254292722812755,
      "grad_norm": 0.10986848920583725,
      "learning_rate": 8.525694732656528e-05,
      "loss": 1.3543,
      "mean_token_accuracy": 0.7478723704814911,
      "num_tokens": 1140503723.0,
      "step": 15870
    },
    {
      "entropy": 0.9184869810938835,
      "epoch": 4.328154810575088,
      "grad_norm": 0.10071445256471634,
      "learning_rate": 8.523563760956274e-05,
      "loss": 1.3477,
      "mean_token_accuracy": 0.7526082947850228,
      "num_tokens": 1141233297.0,
      "step": 15880
    },
    {
      "entropy": 0.9309237048029899,
      "epoch": 4.330880348868901,
      "grad_norm": 0.11324868351221085,
      "learning_rate": 8.521431517039975e-05,
      "loss": 1.3621,
      "mean_token_accuracy": 0.7441210955381393,
      "num_tokens": 1141946146.0,
      "step": 15890
    },
    {
      "entropy": 0.9304992854595184,
      "epoch": 4.333605887162714,
      "grad_norm": 0.10269147157669067,
      "learning_rate": 8.5192980016775e-05,
      "loss": 1.3595,
      "mean_token_accuracy": 0.7481666520237923,
      "num_tokens": 1142670652.0,
      "step": 15900
    },
    {
      "entropy": 0.9177383273839951,
      "epoch": 4.336331425456527,
      "grad_norm": 0.10282835364341736,
      "learning_rate": 8.517163215639173e-05,
      "loss": 1.3519,
      "mean_token_accuracy": 0.7466689467430114,
      "num_tokens": 1143374422.0,
      "step": 15910
    },
    {
      "entropy": 0.9125888720154762,
      "epoch": 4.339056963750341,
      "grad_norm": 0.09721323847770691,
      "learning_rate": 8.515027159695781e-05,
      "loss": 1.3374,
      "mean_token_accuracy": 0.752588203549385,
      "num_tokens": 1144103325.0,
      "step": 15920
    },
    {
      "entropy": 0.9269572705030441,
      "epoch": 4.341782502044154,
      "grad_norm": 0.10542239248752594,
      "learning_rate": 8.512889834618564e-05,
      "loss": 1.3542,
      "mean_token_accuracy": 0.74549660384655,
      "num_tokens": 1144822140.0,
      "step": 15930
    },
    {
      "entropy": 0.9246848747134209,
      "epoch": 4.344508040337967,
      "grad_norm": 0.09933461248874664,
      "learning_rate": 8.510751241179226e-05,
      "loss": 1.3485,
      "mean_token_accuracy": 0.7489146247506142,
      "num_tokens": 1145552344.0,
      "step": 15940
    },
    {
      "entropy": 0.9123001664876937,
      "epoch": 4.34723357863178,
      "grad_norm": 0.10787799954414368,
      "learning_rate": 8.508611380149925e-05,
      "loss": 1.3404,
      "mean_token_accuracy": 0.7499959483742714,
      "num_tokens": 1146273413.0,
      "step": 15950
    },
    {
      "entropy": 0.9135029569268227,
      "epoch": 4.349959116925593,
      "grad_norm": 0.12957337498664856,
      "learning_rate": 8.50647025230328e-05,
      "loss": 1.3445,
      "mean_token_accuracy": 0.752824242413044,
      "num_tokens": 1147009846.0,
      "step": 15960
    },
    {
      "entropy": 0.9256454929709435,
      "epoch": 4.352684655219406,
      "grad_norm": 0.1090337410569191,
      "learning_rate": 8.504327858412364e-05,
      "loss": 1.363,
      "mean_token_accuracy": 0.7460663363337516,
      "num_tokens": 1147720111.0,
      "step": 15970
    },
    {
      "entropy": 0.9286524310708046,
      "epoch": 4.355410193513219,
      "grad_norm": 0.10183461755514145,
      "learning_rate": 8.502184199250709e-05,
      "loss": 1.3548,
      "mean_token_accuracy": 0.7498515471816063,
      "num_tokens": 1148447409.0,
      "step": 15980
    },
    {
      "entropy": 0.9260925754904747,
      "epoch": 4.358135731807032,
      "grad_norm": 0.0972822979092598,
      "learning_rate": 8.500039275592302e-05,
      "loss": 1.3582,
      "mean_token_accuracy": 0.7439198985695838,
      "num_tokens": 1149161319.0,
      "step": 15990
    },
    {
      "entropy": 0.9181397452950477,
      "epoch": 4.360861270100845,
      "grad_norm": 0.08983498066663742,
      "learning_rate": 8.497893088211591e-05,
      "loss": 1.3525,
      "mean_token_accuracy": 0.7522954598069191,
      "num_tokens": 1149891062.0,
      "step": 16000
    },
    {
      "entropy": 0.9218356519937515,
      "epoch": 4.363586808394658,
      "grad_norm": 0.13123436272144318,
      "learning_rate": 8.495745637883474e-05,
      "loss": 1.36,
      "mean_token_accuracy": 0.7450639083981514,
      "num_tokens": 1150596816.0,
      "step": 16010
    },
    {
      "entropy": 0.9280581161379814,
      "epoch": 4.366312346688471,
      "grad_norm": 0.10914012789726257,
      "learning_rate": 8.493596925383309e-05,
      "loss": 1.3494,
      "mean_token_accuracy": 0.7427716001868248,
      "num_tokens": 1151303685.0,
      "step": 16020
    },
    {
      "entropy": 0.9174679771065712,
      "epoch": 4.369037884982284,
      "grad_norm": 0.10081352293491364,
      "learning_rate": 8.491446951486912e-05,
      "loss": 1.344,
      "mean_token_accuracy": 0.7507496654987336,
      "num_tokens": 1152024972.0,
      "step": 16030
    },
    {
      "entropy": 0.9233581557869911,
      "epoch": 4.371763423276097,
      "grad_norm": 0.11699751764535904,
      "learning_rate": 8.489295716970549e-05,
      "loss": 1.3489,
      "mean_token_accuracy": 0.7476333141326904,
      "num_tokens": 1152740442.0,
      "step": 16040
    },
    {
      "entropy": 0.9269112601876259,
      "epoch": 4.37448896156991,
      "grad_norm": 0.10510966181755066,
      "learning_rate": 8.487143222610943e-05,
      "loss": 1.3523,
      "mean_token_accuracy": 0.7471148818731308,
      "num_tokens": 1153457678.0,
      "step": 16050
    },
    {
      "entropy": 0.9161120623350143,
      "epoch": 4.377214499863723,
      "grad_norm": 0.10543408244848251,
      "learning_rate": 8.484989469185274e-05,
      "loss": 1.3473,
      "mean_token_accuracy": 0.7492431044578552,
      "num_tokens": 1154176774.0,
      "step": 16060
    },
    {
      "entropy": 0.929962956905365,
      "epoch": 4.379940038157536,
      "grad_norm": 0.11497306823730469,
      "learning_rate": 8.482834457471172e-05,
      "loss": 1.3654,
      "mean_token_accuracy": 0.744649438560009,
      "num_tokens": 1154886531.0,
      "step": 16070
    },
    {
      "entropy": 0.9178838640451431,
      "epoch": 4.382665576451349,
      "grad_norm": 0.11396227031946182,
      "learning_rate": 8.480678188246729e-05,
      "loss": 1.35,
      "mean_token_accuracy": 0.7467187210917473,
      "num_tokens": 1155596124.0,
      "step": 16080
    },
    {
      "entropy": 0.9303081452846527,
      "epoch": 4.385391114745162,
      "grad_norm": 0.11013922095298767,
      "learning_rate": 8.478520662290483e-05,
      "loss": 1.3514,
      "mean_token_accuracy": 0.7479231730103493,
      "num_tokens": 1156315734.0,
      "step": 16090
    },
    {
      "entropy": 0.9267628148198128,
      "epoch": 4.388116653038975,
      "grad_norm": 0.10404833406209946,
      "learning_rate": 8.476361880381432e-05,
      "loss": 1.3501,
      "mean_token_accuracy": 0.7476871356368064,
      "num_tokens": 1157041302.0,
      "step": 16100
    },
    {
      "entropy": 0.9194824427366257,
      "epoch": 4.390842191332788,
      "grad_norm": 0.10547439754009247,
      "learning_rate": 8.47420184329902e-05,
      "loss": 1.3568,
      "mean_token_accuracy": 0.7509753480553627,
      "num_tokens": 1157759040.0,
      "step": 16110
    },
    {
      "entropy": 0.9184313416481018,
      "epoch": 4.393567729626601,
      "grad_norm": 0.11044033616781235,
      "learning_rate": 8.472040551823153e-05,
      "loss": 1.3452,
      "mean_token_accuracy": 0.7501204147934913,
      "num_tokens": 1158489685.0,
      "step": 16120
    },
    {
      "entropy": 0.922398003935814,
      "epoch": 4.396293267920415,
      "grad_norm": 0.10849247872829437,
      "learning_rate": 8.469878006734185e-05,
      "loss": 1.3504,
      "mean_token_accuracy": 0.748694671690464,
      "num_tokens": 1159215721.0,
      "step": 16130
    },
    {
      "entropy": 0.9304196760058403,
      "epoch": 4.399018806214228,
      "grad_norm": 0.10773295164108276,
      "learning_rate": 8.467714208812923e-05,
      "loss": 1.3597,
      "mean_token_accuracy": 0.7436440542340279,
      "num_tokens": 1159928115.0,
      "step": 16140
    },
    {
      "entropy": 0.9251463919878006,
      "epoch": 4.401744344508041,
      "grad_norm": 0.09861241281032562,
      "learning_rate": 8.465549158840626e-05,
      "loss": 1.3582,
      "mean_token_accuracy": 0.742143839597702,
      "num_tokens": 1160630520.0,
      "step": 16150
    },
    {
      "entropy": 0.922368374466896,
      "epoch": 4.404469882801854,
      "grad_norm": 0.11731061339378357,
      "learning_rate": 8.463382857599007e-05,
      "loss": 1.3534,
      "mean_token_accuracy": 0.7475259497761726,
      "num_tokens": 1161345668.0,
      "step": 16160
    },
    {
      "entropy": 0.9281652629375458,
      "epoch": 4.4071954210956665,
      "grad_norm": 0.11117558926343918,
      "learning_rate": 8.461215305870226e-05,
      "loss": 1.3581,
      "mean_token_accuracy": 0.747750635445118,
      "num_tokens": 1162066131.0,
      "step": 16170
    },
    {
      "entropy": 0.9134318038821221,
      "epoch": 4.4099209593894795,
      "grad_norm": 0.09772191941738129,
      "learning_rate": 8.459046504436905e-05,
      "loss": 1.3446,
      "mean_token_accuracy": 0.7544653728604317,
      "num_tokens": 1162798226.0,
      "step": 16180
    },
    {
      "entropy": 0.9234729900956153,
      "epoch": 4.412646497683292,
      "grad_norm": 0.1031123474240303,
      "learning_rate": 8.456876454082103e-05,
      "loss": 1.3523,
      "mean_token_accuracy": 0.7496284931898117,
      "num_tokens": 1163521418.0,
      "step": 16190
    },
    {
      "entropy": 0.915256017446518,
      "epoch": 4.415372035977105,
      "grad_norm": 0.12567292153835297,
      "learning_rate": 8.454705155589341e-05,
      "loss": 1.34,
      "mean_token_accuracy": 0.7480539917945862,
      "num_tokens": 1164233913.0,
      "step": 16200
    },
    {
      "entropy": 0.9199686855077743,
      "epoch": 4.418097574270918,
      "grad_norm": 0.10363832116127014,
      "learning_rate": 8.45253260974259e-05,
      "loss": 1.3531,
      "mean_token_accuracy": 0.7477450609207154,
      "num_tokens": 1164956517.0,
      "step": 16210
    },
    {
      "entropy": 0.9239618211984635,
      "epoch": 4.420823112564731,
      "grad_norm": 0.09600310772657394,
      "learning_rate": 8.45035881732626e-05,
      "loss": 1.3555,
      "mean_token_accuracy": 0.7452525213360787,
      "num_tokens": 1165671996.0,
      "step": 16220
    },
    {
      "entropy": 0.9234343186020851,
      "epoch": 4.423548650858544,
      "grad_norm": 0.10287973284721375,
      "learning_rate": 8.448183779125227e-05,
      "loss": 1.3539,
      "mean_token_accuracy": 0.7488687202334404,
      "num_tokens": 1166400173.0,
      "step": 16230
    },
    {
      "entropy": 0.9195742890238762,
      "epoch": 4.426274189152357,
      "grad_norm": 0.11038333922624588,
      "learning_rate": 8.446007495924805e-05,
      "loss": 1.3499,
      "mean_token_accuracy": 0.7471941247582435,
      "num_tokens": 1167112628.0,
      "step": 16240
    },
    {
      "entropy": 0.9236080378293992,
      "epoch": 4.428999727446171,
      "grad_norm": 0.10123874247074127,
      "learning_rate": 8.443829968510763e-05,
      "loss": 1.3633,
      "mean_token_accuracy": 0.7433831810951232,
      "num_tokens": 1167817908.0,
      "step": 16250
    },
    {
      "entropy": 0.9259242162108421,
      "epoch": 4.431725265739984,
      "grad_norm": 0.10518823564052582,
      "learning_rate": 8.441651197669316e-05,
      "loss": 1.3514,
      "mean_token_accuracy": 0.7478140458464623,
      "num_tokens": 1168537282.0,
      "step": 16260
    },
    {
      "entropy": 0.9266928464174271,
      "epoch": 4.434450804033797,
      "grad_norm": 0.12261072546243668,
      "learning_rate": 8.439471184187133e-05,
      "loss": 1.3529,
      "mean_token_accuracy": 0.7518995612859726,
      "num_tokens": 1169273826.0,
      "step": 16270
    },
    {
      "entropy": 0.9274758994579315,
      "epoch": 4.43717634232761,
      "grad_norm": 0.10522782057523727,
      "learning_rate": 8.437289928851327e-05,
      "loss": 1.3591,
      "mean_token_accuracy": 0.7453684002161026,
      "num_tokens": 1169989086.0,
      "step": 16280
    },
    {
      "entropy": 0.917501051723957,
      "epoch": 4.439901880621423,
      "grad_norm": 0.10446923226118088,
      "learning_rate": 8.435107432449459e-05,
      "loss": 1.3456,
      "mean_token_accuracy": 0.7490571230649948,
      "num_tokens": 1170715609.0,
      "step": 16290
    },
    {
      "entropy": 0.9176990106701851,
      "epoch": 4.442627418915236,
      "grad_norm": 0.11490264534950256,
      "learning_rate": 8.432923695769543e-05,
      "loss": 1.3471,
      "mean_token_accuracy": 0.7479099735617638,
      "num_tokens": 1171439227.0,
      "step": 16300
    },
    {
      "entropy": 0.9271521344780922,
      "epoch": 4.445352957209049,
      "grad_norm": 0.10181035101413727,
      "learning_rate": 8.430738719600035e-05,
      "loss": 1.3588,
      "mean_token_accuracy": 0.7459352523088455,
      "num_tokens": 1172156027.0,
      "step": 16310
    },
    {
      "entropy": 0.9205224305391312,
      "epoch": 4.448078495502862,
      "grad_norm": 0.14165246486663818,
      "learning_rate": 8.428552504729846e-05,
      "loss": 1.347,
      "mean_token_accuracy": 0.7521676912903785,
      "num_tokens": 1172881377.0,
      "step": 16320
    },
    {
      "entropy": 0.9157086759805679,
      "epoch": 4.4508040337966746,
      "grad_norm": 0.10163746774196625,
      "learning_rate": 8.426365051948324e-05,
      "loss": 1.3508,
      "mean_token_accuracy": 0.7470083206892013,
      "num_tokens": 1173594518.0,
      "step": 16330
    },
    {
      "entropy": 0.9186356648802757,
      "epoch": 4.4535295720904875,
      "grad_norm": 0.10214592516422272,
      "learning_rate": 8.424176362045273e-05,
      "loss": 1.3415,
      "mean_token_accuracy": 0.7489265576004982,
      "num_tokens": 1174314989.0,
      "step": 16340
    },
    {
      "entropy": 0.9227355569601059,
      "epoch": 4.456255110384301,
      "grad_norm": 0.12574359774589539,
      "learning_rate": 8.421986435810938e-05,
      "loss": 1.3541,
      "mean_token_accuracy": 0.7484306424856186,
      "num_tokens": 1175048171.0,
      "step": 16350
    },
    {
      "entropy": 0.9171219453215599,
      "epoch": 4.458980648678114,
      "grad_norm": 0.11906462907791138,
      "learning_rate": 8.419795274036015e-05,
      "loss": 1.3478,
      "mean_token_accuracy": 0.7481898292899132,
      "num_tokens": 1175762446.0,
      "step": 16360
    },
    {
      "entropy": 0.9264452278614044,
      "epoch": 4.461706186971927,
      "grad_norm": 0.10201243311166763,
      "learning_rate": 8.41760287751164e-05,
      "loss": 1.364,
      "mean_token_accuracy": 0.7468571528792382,
      "num_tokens": 1176480854.0,
      "step": 16370
    },
    {
      "entropy": 0.9220310509204864,
      "epoch": 4.46443172526574,
      "grad_norm": 0.10590444505214691,
      "learning_rate": 8.415409247029404e-05,
      "loss": 1.3537,
      "mean_token_accuracy": 0.7467442423105239,
      "num_tokens": 1177190511.0,
      "step": 16380
    },
    {
      "entropy": 0.9257250636816025,
      "epoch": 4.467157263559553,
      "grad_norm": 0.11154093593358994,
      "learning_rate": 8.413214383381336e-05,
      "loss": 1.3552,
      "mean_token_accuracy": 0.7494167581200599,
      "num_tokens": 1177920157.0,
      "step": 16390
    },
    {
      "entropy": 0.9163769647479058,
      "epoch": 4.469882801853366,
      "grad_norm": 0.10316795855760574,
      "learning_rate": 8.411018287359908e-05,
      "loss": 1.3539,
      "mean_token_accuracy": 0.7508632883429527,
      "num_tokens": 1178626392.0,
      "step": 16400
    },
    {
      "entropy": 0.91862373650074,
      "epoch": 4.472608340147179,
      "grad_norm": 0.10479848831892014,
      "learning_rate": 8.408820959758046e-05,
      "loss": 1.3531,
      "mean_token_accuracy": 0.7483569413423539,
      "num_tokens": 1179350119.0,
      "step": 16410
    },
    {
      "entropy": 0.9246098041534424,
      "epoch": 4.475333878440992,
      "grad_norm": 0.10754341632127762,
      "learning_rate": 8.406622401369115e-05,
      "loss": 1.3602,
      "mean_token_accuracy": 0.7459397271275521,
      "num_tokens": 1180061482.0,
      "step": 16420
    },
    {
      "entropy": 0.9221455231308937,
      "epoch": 4.478059416734805,
      "grad_norm": 0.10532528162002563,
      "learning_rate": 8.404422612986923e-05,
      "loss": 1.3549,
      "mean_token_accuracy": 0.7448075026273727,
      "num_tokens": 1180768341.0,
      "step": 16430
    },
    {
      "entropy": 0.9203606501221657,
      "epoch": 4.480784955028618,
      "grad_norm": 0.11600491404533386,
      "learning_rate": 8.402221595405728e-05,
      "loss": 1.3466,
      "mean_token_accuracy": 0.751039968430996,
      "num_tokens": 1181502453.0,
      "step": 16440
    },
    {
      "entropy": 0.920854963362217,
      "epoch": 4.483510493322431,
      "grad_norm": 0.10823602974414825,
      "learning_rate": 8.400019349420226e-05,
      "loss": 1.3483,
      "mean_token_accuracy": 0.7461565658450127,
      "num_tokens": 1182212543.0,
      "step": 16450
    },
    {
      "entropy": 0.9236090317368507,
      "epoch": 4.486236031616245,
      "grad_norm": 0.11785374581813812,
      "learning_rate": 8.39781587582556e-05,
      "loss": 1.3671,
      "mean_token_accuracy": 0.7461556300520897,
      "num_tokens": 1182917582.0,
      "step": 16460
    },
    {
      "entropy": 0.9163339242339135,
      "epoch": 4.488961569910058,
      "grad_norm": 0.10780960321426392,
      "learning_rate": 8.395611175417313e-05,
      "loss": 1.3526,
      "mean_token_accuracy": 0.7482388079166412,
      "num_tokens": 1183624007.0,
      "step": 16470
    },
    {
      "entropy": 0.9186526373028755,
      "epoch": 4.4916871082038705,
      "grad_norm": 0.10127273947000504,
      "learning_rate": 8.393405248991515e-05,
      "loss": 1.347,
      "mean_token_accuracy": 0.7462329477071762,
      "num_tokens": 1184337014.0,
      "step": 16480
    },
    {
      "entropy": 0.9266805961728096,
      "epoch": 4.4944126464976835,
      "grad_norm": 0.10493501275777817,
      "learning_rate": 8.391198097344634e-05,
      "loss": 1.3524,
      "mean_token_accuracy": 0.7469391375780106,
      "num_tokens": 1185064509.0,
      "step": 16490
    },
    {
      "entropy": 0.9277998894453049,
      "epoch": 4.497138184791496,
      "grad_norm": 0.10381454974412918,
      "learning_rate": 8.388989721273587e-05,
      "loss": 1.3546,
      "mean_token_accuracy": 0.7436658173799515,
      "num_tokens": 1185775371.0,
      "step": 16500
    },
    {
      "entropy": 0.9154150173068046,
      "epoch": 4.499863723085309,
      "grad_norm": 0.10464684665203094,
      "learning_rate": 8.386780121575726e-05,
      "loss": 1.3418,
      "mean_token_accuracy": 0.7563139602541924,
      "num_tokens": 1186525221.0,
      "step": 16510
    },
    {
      "entropy": 0.9189923197031021,
      "epoch": 4.502589261379122,
      "grad_norm": 0.10428628325462341,
      "learning_rate": 8.384569299048847e-05,
      "loss": 1.3489,
      "mean_token_accuracy": 0.7505606949329376,
      "num_tokens": 1187263881.0,
      "step": 16520
    },
    {
      "entropy": 0.9158635541796685,
      "epoch": 4.505314799672935,
      "grad_norm": 0.10386273264884949,
      "learning_rate": 8.382357254491193e-05,
      "loss": 1.3484,
      "mean_token_accuracy": 0.7489606320858002,
      "num_tokens": 1187982522.0,
      "step": 16530
    },
    {
      "entropy": 0.9306271061301231,
      "epoch": 4.508040337966748,
      "grad_norm": 0.1186213418841362,
      "learning_rate": 8.38014398870144e-05,
      "loss": 1.3567,
      "mean_token_accuracy": 0.7450304880738259,
      "num_tokens": 1188695919.0,
      "step": 16540
    },
    {
      "entropy": 0.925073404610157,
      "epoch": 4.510765876260561,
      "grad_norm": 0.10341156274080276,
      "learning_rate": 8.377929502478707e-05,
      "loss": 1.3534,
      "mean_token_accuracy": 0.7464147448539734,
      "num_tokens": 1189414411.0,
      "step": 16550
    },
    {
      "entropy": 0.9185432195663452,
      "epoch": 4.513491414554375,
      "grad_norm": 0.10422544181346893,
      "learning_rate": 8.37571379662256e-05,
      "loss": 1.3548,
      "mean_token_accuracy": 0.7488935604691506,
      "num_tokens": 1190132422.0,
      "step": 16560
    },
    {
      "entropy": 0.9278983324766159,
      "epoch": 4.516216952848188,
      "grad_norm": 0.10178219527006149,
      "learning_rate": 8.373496871932998e-05,
      "loss": 1.3594,
      "mean_token_accuracy": 0.7470879390835762,
      "num_tokens": 1190857831.0,
      "step": 16570
    },
    {
      "entropy": 0.9227214083075523,
      "epoch": 4.518942491142001,
      "grad_norm": 0.1024470180273056,
      "learning_rate": 8.371278729210461e-05,
      "loss": 1.3568,
      "mean_token_accuracy": 0.7462427109479904,
      "num_tokens": 1191571574.0,
      "step": 16580
    },
    {
      "entropy": 0.9245967090129852,
      "epoch": 4.521668029435814,
      "grad_norm": 0.11787084490060806,
      "learning_rate": 8.369059369255834e-05,
      "loss": 1.3526,
      "mean_token_accuracy": 0.7433294966816902,
      "num_tokens": 1192267504.0,
      "step": 16590
    },
    {
      "entropy": 0.9241783767938614,
      "epoch": 4.524393567729627,
      "grad_norm": 0.10807237029075623,
      "learning_rate": 8.366838792870435e-05,
      "loss": 1.3475,
      "mean_token_accuracy": 0.7497042864561081,
      "num_tokens": 1193000458.0,
      "step": 16600
    },
    {
      "entropy": 0.9186954587697983,
      "epoch": 4.52711910602344,
      "grad_norm": 0.10416384041309357,
      "learning_rate": 8.364617000856025e-05,
      "loss": 1.3426,
      "mean_token_accuracy": 0.7473282679915428,
      "num_tokens": 1193723650.0,
      "step": 16610
    },
    {
      "entropy": 0.9211036786437035,
      "epoch": 4.529844644317253,
      "grad_norm": 0.107343390583992,
      "learning_rate": 8.362393994014805e-05,
      "loss": 1.3493,
      "mean_token_accuracy": 0.7486666798591614,
      "num_tokens": 1194436374.0,
      "step": 16620
    },
    {
      "entropy": 0.9238960713148117,
      "epoch": 4.532570182611066,
      "grad_norm": 0.10108114033937454,
      "learning_rate": 8.360169773149408e-05,
      "loss": 1.348,
      "mean_token_accuracy": 0.7451175212860107,
      "num_tokens": 1195156490.0,
      "step": 16630
    },
    {
      "entropy": 0.9179178893566131,
      "epoch": 4.5352957209048785,
      "grad_norm": 0.09835729002952576,
      "learning_rate": 8.357944339062915e-05,
      "loss": 1.3473,
      "mean_token_accuracy": 0.7542638510465622,
      "num_tokens": 1195902942.0,
      "step": 16640
    },
    {
      "entropy": 0.9226110145449639,
      "epoch": 4.5380212591986915,
      "grad_norm": 0.10751954466104507,
      "learning_rate": 8.355717692558837e-05,
      "loss": 1.3537,
      "mean_token_accuracy": 0.7422856971621513,
      "num_tokens": 1196601758.0,
      "step": 16650
    },
    {
      "entropy": 0.9215986758470536,
      "epoch": 4.540746797492504,
      "grad_norm": 0.0961076095700264,
      "learning_rate": 8.353489834441129e-05,
      "loss": 1.3435,
      "mean_token_accuracy": 0.7530729457736015,
      "num_tokens": 1197337952.0,
      "step": 16660
    },
    {
      "entropy": 0.9258883237838745,
      "epoch": 4.543472335786317,
      "grad_norm": 0.10355409979820251,
      "learning_rate": 8.351260765514175e-05,
      "loss": 1.3663,
      "mean_token_accuracy": 0.742791873216629,
      "num_tokens": 1198035431.0,
      "step": 16670
    },
    {
      "entropy": 0.9232835918664932,
      "epoch": 4.546197874080131,
      "grad_norm": 0.1080857664346695,
      "learning_rate": 8.349030486582805e-05,
      "loss": 1.3494,
      "mean_token_accuracy": 0.7489354580640792,
      "num_tokens": 1198753666.0,
      "step": 16680
    },
    {
      "entropy": 0.9208525791764259,
      "epoch": 4.548923412373944,
      "grad_norm": 0.1087987944483757,
      "learning_rate": 8.346798998452282e-05,
      "loss": 1.3547,
      "mean_token_accuracy": 0.7472927644848824,
      "num_tokens": 1199471906.0,
      "step": 16690
    },
    {
      "entropy": 0.9178042456507682,
      "epoch": 4.551648950667757,
      "grad_norm": 0.1054878681898117,
      "learning_rate": 8.344566301928306e-05,
      "loss": 1.3456,
      "mean_token_accuracy": 0.7514221504330635,
      "num_tokens": 1200189269.0,
      "step": 16700
    },
    {
      "entropy": 0.9227500975131988,
      "epoch": 4.55437448896157,
      "grad_norm": 0.10009706020355225,
      "learning_rate": 8.342332397817011e-05,
      "loss": 1.3527,
      "mean_token_accuracy": 0.7442700415849686,
      "num_tokens": 1200904850.0,
      "step": 16710
    },
    {
      "entropy": 0.9224673196673393,
      "epoch": 4.557100027255383,
      "grad_norm": 0.12601953744888306,
      "learning_rate": 8.340097286924975e-05,
      "loss": 1.3479,
      "mean_token_accuracy": 0.748690877854824,
      "num_tokens": 1201625428.0,
      "step": 16720
    },
    {
      "entropy": 0.9233771145343781,
      "epoch": 4.559825565549196,
      "grad_norm": 0.10821715742349625,
      "learning_rate": 8.337860970059199e-05,
      "loss": 1.3532,
      "mean_token_accuracy": 0.7455072551965714,
      "num_tokens": 1202341734.0,
      "step": 16730
    },
    {
      "entropy": 0.9218176856637001,
      "epoch": 4.562551103843009,
      "grad_norm": 0.10075296461582184,
      "learning_rate": 8.33562344802713e-05,
      "loss": 1.3556,
      "mean_token_accuracy": 0.7500150710344314,
      "num_tokens": 1203076540.0,
      "step": 16740
    },
    {
      "entropy": 0.9333092331886291,
      "epoch": 4.565276642136822,
      "grad_norm": 0.10132449865341187,
      "learning_rate": 8.333384721636646e-05,
      "loss": 1.3603,
      "mean_token_accuracy": 0.7441597357392311,
      "num_tokens": 1203792960.0,
      "step": 16750
    },
    {
      "entropy": 0.9210795417428017,
      "epoch": 4.568002180430635,
      "grad_norm": 0.09611700475215912,
      "learning_rate": 8.331144791696059e-05,
      "loss": 1.3486,
      "mean_token_accuracy": 0.7507175087928772,
      "num_tokens": 1204522218.0,
      "step": 16760
    },
    {
      "entropy": 0.9296272456645965,
      "epoch": 4.570727718724448,
      "grad_norm": 0.10803563892841339,
      "learning_rate": 8.32890365901412e-05,
      "loss": 1.3538,
      "mean_token_accuracy": 0.7462391838431358,
      "num_tokens": 1205239556.0,
      "step": 16770
    },
    {
      "entropy": 0.9215833842754364,
      "epoch": 4.5734532570182616,
      "grad_norm": 0.10197755694389343,
      "learning_rate": 8.326661324400008e-05,
      "loss": 1.3552,
      "mean_token_accuracy": 0.7484742417931557,
      "num_tokens": 1205955974.0,
      "step": 16780
    },
    {
      "entropy": 0.9258045464754104,
      "epoch": 4.5761787953120745,
      "grad_norm": 0.10073843598365784,
      "learning_rate": 8.32441778866334e-05,
      "loss": 1.3617,
      "mean_token_accuracy": 0.7447678685188294,
      "num_tokens": 1206664924.0,
      "step": 16790
    },
    {
      "entropy": 0.9241404712200165,
      "epoch": 4.5789043336058874,
      "grad_norm": 0.10945308208465576,
      "learning_rate": 8.322173052614169e-05,
      "loss": 1.3563,
      "mean_token_accuracy": 0.7472848266363143,
      "num_tokens": 1207376945.0,
      "step": 16800
    },
    {
      "entropy": 0.9194119274616241,
      "epoch": 4.5816298718997,
      "grad_norm": 0.10351152718067169,
      "learning_rate": 8.319927117062973e-05,
      "loss": 1.3489,
      "mean_token_accuracy": 0.7491079226136208,
      "num_tokens": 1208104503.0,
      "step": 16810
    },
    {
      "entropy": 0.9161817282438278,
      "epoch": 4.584355410193513,
      "grad_norm": 0.13375310599803925,
      "learning_rate": 8.317679982820671e-05,
      "loss": 1.3541,
      "mean_token_accuracy": 0.752979950606823,
      "num_tokens": 1208833481.0,
      "step": 16820
    },
    {
      "entropy": 0.9225863084197045,
      "epoch": 4.587080948487326,
      "grad_norm": 0.10402826964855194,
      "learning_rate": 8.315431650698613e-05,
      "loss": 1.3586,
      "mean_token_accuracy": 0.7458830237388611,
      "num_tokens": 1209547234.0,
      "step": 16830
    },
    {
      "entropy": 0.9251091539859772,
      "epoch": 4.589806486781139,
      "grad_norm": 0.10585848242044449,
      "learning_rate": 8.313182121508578e-05,
      "loss": 1.3532,
      "mean_token_accuracy": 0.7482419028878212,
      "num_tokens": 1210267468.0,
      "step": 16840
    },
    {
      "entropy": 0.9181981027126312,
      "epoch": 4.592532025074952,
      "grad_norm": 0.10736094415187836,
      "learning_rate": 8.310931396062782e-05,
      "loss": 1.35,
      "mean_token_accuracy": 0.7476749077439309,
      "num_tokens": 1210981391.0,
      "step": 16850
    },
    {
      "entropy": 0.9191288724541664,
      "epoch": 4.595257563368765,
      "grad_norm": 0.11005821079015732,
      "learning_rate": 8.30867947517387e-05,
      "loss": 1.3475,
      "mean_token_accuracy": 0.7482886627316475,
      "num_tokens": 1211699743.0,
      "step": 16860
    },
    {
      "entropy": 0.9240297868847847,
      "epoch": 4.597983101662578,
      "grad_norm": 0.10537801682949066,
      "learning_rate": 8.306426359654917e-05,
      "loss": 1.357,
      "mean_token_accuracy": 0.7482584595680237,
      "num_tokens": 1212402209.0,
      "step": 16870
    },
    {
      "entropy": 0.9225147813558578,
      "epoch": 4.600708639956391,
      "grad_norm": 0.11033676564693451,
      "learning_rate": 8.304172050319436e-05,
      "loss": 1.3596,
      "mean_token_accuracy": 0.7457825124263764,
      "num_tokens": 1213118934.0,
      "step": 16880
    },
    {
      "entropy": 0.9162515133619309,
      "epoch": 4.603434178250204,
      "grad_norm": 0.12636755406856537,
      "learning_rate": 8.301916547981365e-05,
      "loss": 1.3517,
      "mean_token_accuracy": 0.7530664309859276,
      "num_tokens": 1213843129.0,
      "step": 16890
    },
    {
      "entropy": 0.9205801919102669,
      "epoch": 4.606159716544018,
      "grad_norm": 0.11650549620389938,
      "learning_rate": 8.299659853455074e-05,
      "loss": 1.354,
      "mean_token_accuracy": 0.7426133722066879,
      "num_tokens": 1214554615.0,
      "step": 16900
    },
    {
      "entropy": 0.9172195091843605,
      "epoch": 4.608885254837831,
      "grad_norm": 0.09601568430662155,
      "learning_rate": 8.297401967555362e-05,
      "loss": 1.3469,
      "mean_token_accuracy": 0.7458287104964256,
      "num_tokens": 1215263942.0,
      "step": 16910
    },
    {
      "entropy": 0.9214783206582069,
      "epoch": 4.611610793131644,
      "grad_norm": 0.10365541279315948,
      "learning_rate": 8.295142891097466e-05,
      "loss": 1.3524,
      "mean_token_accuracy": 0.7539938539266586,
      "num_tokens": 1216001651.0,
      "step": 16920
    },
    {
      "entropy": 0.9217340096831321,
      "epoch": 4.614336331425457,
      "grad_norm": 0.10575239360332489,
      "learning_rate": 8.292882624897041e-05,
      "loss": 1.3543,
      "mean_token_accuracy": 0.7463909596204757,
      "num_tokens": 1216711332.0,
      "step": 16930
    },
    {
      "entropy": 0.9175868049263954,
      "epoch": 4.61706186971927,
      "grad_norm": 0.11438553035259247,
      "learning_rate": 8.290621169770179e-05,
      "loss": 1.3477,
      "mean_token_accuracy": 0.7449107185006142,
      "num_tokens": 1217408569.0,
      "step": 16940
    },
    {
      "entropy": 0.916756784915924,
      "epoch": 4.6197874080130825,
      "grad_norm": 0.10522891581058502,
      "learning_rate": 8.288358526533401e-05,
      "loss": 1.3495,
      "mean_token_accuracy": 0.7468285739421845,
      "num_tokens": 1218121983.0,
      "step": 16950
    },
    {
      "entropy": 0.9158628568053245,
      "epoch": 4.6225129463068955,
      "grad_norm": 0.1695455014705658,
      "learning_rate": 8.286094696003655e-05,
      "loss": 1.3422,
      "mean_token_accuracy": 0.7501478627324104,
      "num_tokens": 1218844701.0,
      "step": 16960
    },
    {
      "entropy": 0.9222366034984588,
      "epoch": 4.625238484600708,
      "grad_norm": 0.12280505150556564,
      "learning_rate": 8.283829678998318e-05,
      "loss": 1.348,
      "mean_token_accuracy": 0.7487332820892334,
      "num_tokens": 1219555849.0,
      "step": 16970
    },
    {
      "entropy": 0.9117548048496247,
      "epoch": 4.627964022894521,
      "grad_norm": 0.09231628477573395,
      "learning_rate": 8.281563476335197e-05,
      "loss": 1.3357,
      "mean_token_accuracy": 0.751614136993885,
      "num_tokens": 1220291533.0,
      "step": 16980
    },
    {
      "entropy": 0.9225894793868065,
      "epoch": 4.630689561188335,
      "grad_norm": 0.109446220099926,
      "learning_rate": 8.279296088832525e-05,
      "loss": 1.3495,
      "mean_token_accuracy": 0.7509663596749305,
      "num_tokens": 1221016163.0,
      "step": 16990
    },
    {
      "entropy": 0.9290517643094063,
      "epoch": 4.633415099482148,
      "grad_norm": 0.10743549466133118,
      "learning_rate": 8.277027517308965e-05,
      "loss": 1.3533,
      "mean_token_accuracy": 0.7469459012150764,
      "num_tokens": 1221746578.0,
      "step": 17000
    },
    {
      "entropy": 0.9262382388114929,
      "epoch": 4.636140637775961,
      "grad_norm": 0.10961365699768066,
      "learning_rate": 8.274757762583605e-05,
      "loss": 1.3579,
      "mean_token_accuracy": 0.7437677323818207,
      "num_tokens": 1222447891.0,
      "step": 17010
    },
    {
      "entropy": 0.9256102323532105,
      "epoch": 4.638866176069774,
      "grad_norm": 0.17248821258544922,
      "learning_rate": 8.27248682547596e-05,
      "loss": 1.3548,
      "mean_token_accuracy": 0.7464618891477585,
      "num_tokens": 1223168990.0,
      "step": 17020
    },
    {
      "entropy": 0.9164867550134659,
      "epoch": 4.641591714363587,
      "grad_norm": 0.11628459393978119,
      "learning_rate": 8.270214706805975e-05,
      "loss": 1.3497,
      "mean_token_accuracy": 0.7500196889042854,
      "num_tokens": 1223888373.0,
      "step": 17030
    },
    {
      "entropy": 0.9233259737491608,
      "epoch": 4.6443172526574,
      "grad_norm": 0.10588211566209793,
      "learning_rate": 8.267941407394021e-05,
      "loss": 1.3546,
      "mean_token_accuracy": 0.7507358610630035,
      "num_tokens": 1224617523.0,
      "step": 17040
    },
    {
      "entropy": 0.9266725420951843,
      "epoch": 4.647042790951213,
      "grad_norm": 0.09244404733181,
      "learning_rate": 8.265666928060894e-05,
      "loss": 1.3634,
      "mean_token_accuracy": 0.7496835634112358,
      "num_tokens": 1225339096.0,
      "step": 17050
    },
    {
      "entropy": 0.9150111705064774,
      "epoch": 4.649768329245026,
      "grad_norm": 0.10257738083600998,
      "learning_rate": 8.263391269627814e-05,
      "loss": 1.3442,
      "mean_token_accuracy": 0.7501561403274536,
      "num_tokens": 1226060970.0,
      "step": 17060
    },
    {
      "entropy": 0.9162648782134056,
      "epoch": 4.652493867538839,
      "grad_norm": 0.10511647909879684,
      "learning_rate": 8.261114432916433e-05,
      "loss": 1.3461,
      "mean_token_accuracy": 0.7498621806502342,
      "num_tokens": 1226780834.0,
      "step": 17070
    },
    {
      "entropy": 0.9222771629691124,
      "epoch": 4.655219405832652,
      "grad_norm": 0.10543674975633621,
      "learning_rate": 8.25883641874882e-05,
      "loss": 1.3545,
      "mean_token_accuracy": 0.7483288168907165,
      "num_tokens": 1227494220.0,
      "step": 17080
    },
    {
      "entropy": 0.9159112647175789,
      "epoch": 4.657944944126465,
      "grad_norm": 0.1046845093369484,
      "learning_rate": 8.256557227947475e-05,
      "loss": 1.3393,
      "mean_token_accuracy": 0.751229789853096,
      "num_tokens": 1228228148.0,
      "step": 17090
    },
    {
      "entropy": 0.9199364632368088,
      "epoch": 4.660670482420278,
      "grad_norm": 0.10549994558095932,
      "learning_rate": 8.254276861335322e-05,
      "loss": 1.3437,
      "mean_token_accuracy": 0.7498672157526016,
      "num_tokens": 1228954174.0,
      "step": 17100
    },
    {
      "entropy": 0.9175642892718315,
      "epoch": 4.6633960207140905,
      "grad_norm": 0.09940032660961151,
      "learning_rate": 8.25199531973571e-05,
      "loss": 1.348,
      "mean_token_accuracy": 0.7468050643801689,
      "num_tokens": 1229667994.0,
      "step": 17110
    },
    {
      "entropy": 0.912084773182869,
      "epoch": 4.666121559007904,
      "grad_norm": 0.10739535838365555,
      "learning_rate": 8.249712603972409e-05,
      "loss": 1.3437,
      "mean_token_accuracy": 0.7519252315163613,
      "num_tokens": 1230389211.0,
      "step": 17120
    },
    {
      "entropy": 0.9226089671254158,
      "epoch": 4.668847097301717,
      "grad_norm": 0.09731782227754593,
      "learning_rate": 8.247428714869616e-05,
      "loss": 1.3479,
      "mean_token_accuracy": 0.74716587215662,
      "num_tokens": 1231097082.0,
      "step": 17130
    },
    {
      "entropy": 0.9198968097567558,
      "epoch": 4.67157263559553,
      "grad_norm": 0.10072658210992813,
      "learning_rate": 8.245143653251952e-05,
      "loss": 1.3422,
      "mean_token_accuracy": 0.7485209181904793,
      "num_tokens": 1231808943.0,
      "step": 17140
    },
    {
      "entropy": 0.9217645287513733,
      "epoch": 4.674298173889343,
      "grad_norm": 0.09910709410905838,
      "learning_rate": 8.242857419944459e-05,
      "loss": 1.3496,
      "mean_token_accuracy": 0.7481761828064919,
      "num_tokens": 1232528394.0,
      "step": 17150
    },
    {
      "entropy": 0.9187888920307159,
      "epoch": 4.677023712183156,
      "grad_norm": 0.1049736887216568,
      "learning_rate": 8.240570015772604e-05,
      "loss": 1.3486,
      "mean_token_accuracy": 0.7452646285295487,
      "num_tokens": 1233236436.0,
      "step": 17160
    },
    {
      "entropy": 0.9299355641007423,
      "epoch": 4.679749250476969,
      "grad_norm": 0.10277475416660309,
      "learning_rate": 8.238281441562273e-05,
      "loss": 1.3641,
      "mean_token_accuracy": 0.7394877701997757,
      "num_tokens": 1233939860.0,
      "step": 17170
    },
    {
      "entropy": 0.9240302801132202,
      "epoch": 4.682474788770782,
      "grad_norm": 0.1163918524980545,
      "learning_rate": 8.235991698139783e-05,
      "loss": 1.3528,
      "mean_token_accuracy": 0.7462989240884781,
      "num_tokens": 1234663102.0,
      "step": 17180
    },
    {
      "entropy": 0.9209274545311927,
      "epoch": 4.685200327064595,
      "grad_norm": 0.10490309447050095,
      "learning_rate": 8.23370078633186e-05,
      "loss": 1.3527,
      "mean_token_accuracy": 0.7471803933382034,
      "num_tokens": 1235386239.0,
      "step": 17190
    },
    {
      "entropy": 0.928577084839344,
      "epoch": 4.687925865358408,
      "grad_norm": 0.11143554747104645,
      "learning_rate": 8.231408706965666e-05,
      "loss": 1.3562,
      "mean_token_accuracy": 0.7434001788496971,
      "num_tokens": 1236088357.0,
      "step": 17200
    },
    {
      "entropy": 0.9305694311857223,
      "epoch": 4.690651403652222,
      "grad_norm": 0.10430223494768143,
      "learning_rate": 8.229115460868775e-05,
      "loss": 1.3588,
      "mean_token_accuracy": 0.7478098511695862,
      "num_tokens": 1236814332.0,
      "step": 17210
    },
    {
      "entropy": 0.9143865138292313,
      "epoch": 4.693376941946035,
      "grad_norm": 0.09553319960832596,
      "learning_rate": 8.226821048869184e-05,
      "loss": 1.3393,
      "mean_token_accuracy": 0.7473413035273552,
      "num_tokens": 1237528620.0,
      "step": 17220
    },
    {
      "entropy": 0.916749370098114,
      "epoch": 4.696102480239848,
      "grad_norm": 0.09525775164365768,
      "learning_rate": 8.224525471795317e-05,
      "loss": 1.3455,
      "mean_token_accuracy": 0.7485346272587776,
      "num_tokens": 1238249386.0,
      "step": 17230
    },
    {
      "entropy": 0.921876285970211,
      "epoch": 4.698828018533661,
      "grad_norm": 0.10387054085731506,
      "learning_rate": 8.222228730476009e-05,
      "loss": 1.3567,
      "mean_token_accuracy": 0.7456609532237053,
      "num_tokens": 1238960045.0,
      "step": 17240
    },
    {
      "entropy": 0.919804036617279,
      "epoch": 4.701553556827474,
      "grad_norm": 0.11917005479335785,
      "learning_rate": 8.21993082574052e-05,
      "loss": 1.3517,
      "mean_token_accuracy": 0.7492782562971115,
      "num_tokens": 1239684245.0,
      "step": 17250
    },
    {
      "entropy": 0.926070973277092,
      "epoch": 4.7042790951212865,
      "grad_norm": 0.11269728094339371,
      "learning_rate": 8.217631758418534e-05,
      "loss": 1.3528,
      "mean_token_accuracy": 0.7491124123334885,
      "num_tokens": 1240409991.0,
      "step": 17260
    },
    {
      "entropy": 0.9213026985526085,
      "epoch": 4.7070046334150994,
      "grad_norm": 0.10869846493005753,
      "learning_rate": 8.21533152934015e-05,
      "loss": 1.3497,
      "mean_token_accuracy": 0.7475429430603981,
      "num_tokens": 1241122725.0,
      "step": 17270
    },
    {
      "entropy": 0.9191406026482583,
      "epoch": 4.709730171708912,
      "grad_norm": 0.1102442592382431,
      "learning_rate": 8.213030139335885e-05,
      "loss": 1.3506,
      "mean_token_accuracy": 0.7509415715932846,
      "num_tokens": 1241837856.0,
      "step": 17280
    },
    {
      "entropy": 0.9176264256238937,
      "epoch": 4.712455710002725,
      "grad_norm": 0.11132961511611938,
      "learning_rate": 8.210727589236679e-05,
      "loss": 1.36,
      "mean_token_accuracy": 0.7469341784715653,
      "num_tokens": 1242542396.0,
      "step": 17290
    },
    {
      "entropy": 0.9260061651468277,
      "epoch": 4.715181248296538,
      "grad_norm": 0.12213858962059021,
      "learning_rate": 8.208423879873892e-05,
      "loss": 1.3524,
      "mean_token_accuracy": 0.747686205804348,
      "num_tokens": 1243266256.0,
      "step": 17300
    },
    {
      "entropy": 0.9204802647233009,
      "epoch": 4.717906786590351,
      "grad_norm": 0.10788416117429733,
      "learning_rate": 8.206119012079297e-05,
      "loss": 1.3562,
      "mean_token_accuracy": 0.7475770086050033,
      "num_tokens": 1243985897.0,
      "step": 17310
    },
    {
      "entropy": 0.9188668221235275,
      "epoch": 4.720632324884164,
      "grad_norm": 0.1071867048740387,
      "learning_rate": 8.203812986685088e-05,
      "loss": 1.3394,
      "mean_token_accuracy": 0.749122953414917,
      "num_tokens": 1244707866.0,
      "step": 17320
    },
    {
      "entropy": 0.9235012277960777,
      "epoch": 4.723357863177978,
      "grad_norm": 0.10762520134449005,
      "learning_rate": 8.201505804523879e-05,
      "loss": 1.3561,
      "mean_token_accuracy": 0.7462771251797676,
      "num_tokens": 1245429738.0,
      "step": 17330
    },
    {
      "entropy": 0.9198080718517303,
      "epoch": 4.726083401471791,
      "grad_norm": 0.10488855838775635,
      "learning_rate": 8.199197466428699e-05,
      "loss": 1.3522,
      "mean_token_accuracy": 0.7489576727151871,
      "num_tokens": 1246143218.0,
      "step": 17340
    },
    {
      "entropy": 0.9154806345701217,
      "epoch": 4.728808939765604,
      "grad_norm": 0.10319756716489792,
      "learning_rate": 8.196887973232996e-05,
      "loss": 1.3449,
      "mean_token_accuracy": 0.7500908523797989,
      "num_tokens": 1246862456.0,
      "step": 17350
    },
    {
      "entropy": 0.9225591003894806,
      "epoch": 4.731534478059417,
      "grad_norm": 0.11112985759973526,
      "learning_rate": 8.194577325770634e-05,
      "loss": 1.3521,
      "mean_token_accuracy": 0.7449140757322311,
      "num_tokens": 1247576328.0,
      "step": 17360
    },
    {
      "entropy": 0.9286596611142158,
      "epoch": 4.73426001635323,
      "grad_norm": 0.10459794104099274,
      "learning_rate": 8.192265524875894e-05,
      "loss": 1.3608,
      "mean_token_accuracy": 0.7479562908411026,
      "num_tokens": 1248298945.0,
      "step": 17370
    },
    {
      "entropy": 0.9241178303956985,
      "epoch": 4.736985554647043,
      "grad_norm": 0.1349504292011261,
      "learning_rate": 8.189952571383475e-05,
      "loss": 1.3501,
      "mean_token_accuracy": 0.744536454975605,
      "num_tokens": 1249012614.0,
      "step": 17380
    },
    {
      "entropy": 0.9124541774392128,
      "epoch": 4.739711092940856,
      "grad_norm": 0.10812727361917496,
      "learning_rate": 8.18763846612849e-05,
      "loss": 1.3422,
      "mean_token_accuracy": 0.7497065871953964,
      "num_tokens": 1249711886.0,
      "step": 17390
    },
    {
      "entropy": 0.9234019756317139,
      "epoch": 4.742436631234669,
      "grad_norm": 0.12515738606452942,
      "learning_rate": 8.185323209946465e-05,
      "loss": 1.3523,
      "mean_token_accuracy": 0.7470652341842652,
      "num_tokens": 1250437022.0,
      "step": 17400
    },
    {
      "entropy": 0.9112591460347176,
      "epoch": 4.745162169528482,
      "grad_norm": 0.21318070590496063,
      "learning_rate": 8.18300680367335e-05,
      "loss": 1.3447,
      "mean_token_accuracy": 0.7522909045219421,
      "num_tokens": 1251163833.0,
      "step": 17410
    },
    {
      "entropy": 0.9248941242694855,
      "epoch": 4.7478877078222945,
      "grad_norm": 0.10058721154928207,
      "learning_rate": 8.180689248145503e-05,
      "loss": 1.3605,
      "mean_token_accuracy": 0.7427182897925377,
      "num_tokens": 1251869086.0,
      "step": 17420
    },
    {
      "entropy": 0.9240973278880119,
      "epoch": 4.750613246116108,
      "grad_norm": 0.10814929753541946,
      "learning_rate": 8.178370544199701e-05,
      "loss": 1.3529,
      "mean_token_accuracy": 0.7461601093411445,
      "num_tokens": 1252580434.0,
      "step": 17430
    },
    {
      "entropy": 0.9222966253757476,
      "epoch": 4.753338784409921,
      "grad_norm": 0.11058203130960464,
      "learning_rate": 8.176050692673134e-05,
      "loss": 1.3509,
      "mean_token_accuracy": 0.7479978322982788,
      "num_tokens": 1253310826.0,
      "step": 17440
    },
    {
      "entropy": 0.9145761653780937,
      "epoch": 4.756064322703734,
      "grad_norm": 0.1636810302734375,
      "learning_rate": 8.173729694403404e-05,
      "loss": 1.3468,
      "mean_token_accuracy": 0.7477872595191002,
      "num_tokens": 1254035147.0,
      "step": 17450
    },
    {
      "entropy": 0.9174238905310631,
      "epoch": 4.758789860997547,
      "grad_norm": 0.1089249700307846,
      "learning_rate": 8.171407550228532e-05,
      "loss": 1.3397,
      "mean_token_accuracy": 0.7504594132304192,
      "num_tokens": 1254764538.0,
      "step": 17460
    },
    {
      "entropy": 0.9233099758625031,
      "epoch": 4.76151539929136,
      "grad_norm": 0.10400953143835068,
      "learning_rate": 8.169084260986948e-05,
      "loss": 1.3541,
      "mean_token_accuracy": 0.7493297189474106,
      "num_tokens": 1255487377.0,
      "step": 17470
    },
    {
      "entropy": 0.9232550159096717,
      "epoch": 4.764240937585173,
      "grad_norm": 0.11339503526687622,
      "learning_rate": 8.166759827517498e-05,
      "loss": 1.3514,
      "mean_token_accuracy": 0.7459690898656846,
      "num_tokens": 1256210228.0,
      "step": 17480
    },
    {
      "entropy": 0.9163402050733567,
      "epoch": 4.766966475878986,
      "grad_norm": 0.10250867158174515,
      "learning_rate": 8.164434250659441e-05,
      "loss": 1.3437,
      "mean_token_accuracy": 0.748111541569233,
      "num_tokens": 1256935585.0,
      "step": 17490
    },
    {
      "entropy": 0.9241256102919578,
      "epoch": 4.769692014172799,
      "grad_norm": 0.11915501952171326,
      "learning_rate": 8.16210753125245e-05,
      "loss": 1.3509,
      "mean_token_accuracy": 0.7453058168292046,
      "num_tokens": 1257648523.0,
      "step": 17500
    },
    {
      "entropy": 0.9295717850327492,
      "epoch": 4.772417552466612,
      "grad_norm": 0.1178000271320343,
      "learning_rate": 8.159779670136608e-05,
      "loss": 1.3574,
      "mean_token_accuracy": 0.7509691178798675,
      "num_tokens": 1258375416.0,
      "step": 17510
    },
    {
      "entropy": 0.9127016827464104,
      "epoch": 4.775143090760425,
      "grad_norm": 0.11207820475101471,
      "learning_rate": 8.15745066815241e-05,
      "loss": 1.3438,
      "mean_token_accuracy": 0.7531407192349434,
      "num_tokens": 1259099537.0,
      "step": 17520
    },
    {
      "entropy": 0.924431873857975,
      "epoch": 4.777868629054238,
      "grad_norm": 0.10758596658706665,
      "learning_rate": 8.155120526140766e-05,
      "loss": 1.3475,
      "mean_token_accuracy": 0.7530660226941108,
      "num_tokens": 1259840085.0,
      "step": 17530
    },
    {
      "entropy": 0.9277138888835907,
      "epoch": 4.780594167348051,
      "grad_norm": 0.13582734763622284,
      "learning_rate": 8.152789244942996e-05,
      "loss": 1.3571,
      "mean_token_accuracy": 0.7471134886145592,
      "num_tokens": 1260562448.0,
      "step": 17540
    },
    {
      "entropy": 0.9182221814990044,
      "epoch": 4.783319705641865,
      "grad_norm": 0.1102224588394165,
      "learning_rate": 8.15045682540083e-05,
      "loss": 1.3433,
      "mean_token_accuracy": 0.7517937749624253,
      "num_tokens": 1261289261.0,
      "step": 17550
    },
    {
      "entropy": 0.917127600312233,
      "epoch": 4.7860452439356775,
      "grad_norm": 0.10573709011077881,
      "learning_rate": 8.14812326835641e-05,
      "loss": 1.3528,
      "mean_token_accuracy": 0.757950046658516,
      "num_tokens": 1262033603.0,
      "step": 17560
    },
    {
      "entropy": 0.9341249838471413,
      "epoch": 4.7887707822294905,
      "grad_norm": 0.10214850306510925,
      "learning_rate": 8.145788574652288e-05,
      "loss": 1.3639,
      "mean_token_accuracy": 0.7426125407218933,
      "num_tokens": 1262752998.0,
      "step": 17570
    },
    {
      "entropy": 0.9180441945791245,
      "epoch": 4.791496320523303,
      "grad_norm": 0.1256493777036667,
      "learning_rate": 8.14345274513143e-05,
      "loss": 1.3409,
      "mean_token_accuracy": 0.7528748378157616,
      "num_tokens": 1263483934.0,
      "step": 17580
    },
    {
      "entropy": 0.9169729307293892,
      "epoch": 4.794221858817116,
      "grad_norm": 0.098321832716465,
      "learning_rate": 8.141115780637208e-05,
      "loss": 1.3517,
      "mean_token_accuracy": 0.7468766659498215,
      "num_tokens": 1264203136.0,
      "step": 17590
    },
    {
      "entropy": 0.9212764590978623,
      "epoch": 4.796947397110929,
      "grad_norm": 0.11560304462909698,
      "learning_rate": 8.138777682013403e-05,
      "loss": 1.3473,
      "mean_token_accuracy": 0.7478474527597427,
      "num_tokens": 1264927412.0,
      "step": 17600
    },
    {
      "entropy": 0.916501834988594,
      "epoch": 4.799672935404742,
      "grad_norm": 0.10708596557378769,
      "learning_rate": 8.136438450104212e-05,
      "loss": 1.3471,
      "mean_token_accuracy": 0.7494879648089409,
      "num_tokens": 1265638281.0,
      "step": 17610
    },
    {
      "entropy": 0.9157280281186104,
      "epoch": 4.802398473698555,
      "grad_norm": 0.09304467588663101,
      "learning_rate": 8.134098085754233e-05,
      "loss": 1.346,
      "mean_token_accuracy": 0.7529150784015656,
      "num_tokens": 1266367884.0,
      "step": 17620
    },
    {
      "entropy": 0.9149423062801361,
      "epoch": 4.805124011992368,
      "grad_norm": 0.11460596323013306,
      "learning_rate": 8.131756589808479e-05,
      "loss": 1.3398,
      "mean_token_accuracy": 0.7494656577706337,
      "num_tokens": 1267096886.0,
      "step": 17630
    },
    {
      "entropy": 0.9196400806307793,
      "epoch": 4.807849550286182,
      "grad_norm": 0.09922295808792114,
      "learning_rate": 8.129413963112369e-05,
      "loss": 1.3484,
      "mean_token_accuracy": 0.7489532619714737,
      "num_tokens": 1267815503.0,
      "step": 17640
    },
    {
      "entropy": 0.92071463316679,
      "epoch": 4.810575088579995,
      "grad_norm": 0.12019436806440353,
      "learning_rate": 8.127070206511729e-05,
      "loss": 1.3439,
      "mean_token_accuracy": 0.7510943159461021,
      "num_tokens": 1268547915.0,
      "step": 17650
    },
    {
      "entropy": 0.9236974358558655,
      "epoch": 4.813300626873808,
      "grad_norm": 0.11996635049581528,
      "learning_rate": 8.124725320852796e-05,
      "loss": 1.3582,
      "mean_token_accuracy": 0.748077042400837,
      "num_tokens": 1269267387.0,
      "step": 17660
    },
    {
      "entropy": 0.9122890293598175,
      "epoch": 4.816026165167621,
      "grad_norm": 0.11178936809301376,
      "learning_rate": 8.122379306982213e-05,
      "loss": 1.3329,
      "mean_token_accuracy": 0.7529049053788185,
      "num_tokens": 1270003398.0,
      "step": 17670
    },
    {
      "entropy": 0.9168417230248451,
      "epoch": 4.818751703461434,
      "grad_norm": 0.14724519848823547,
      "learning_rate": 8.12003216574703e-05,
      "loss": 1.3509,
      "mean_token_accuracy": 0.749293364584446,
      "num_tokens": 1270718938.0,
      "step": 17680
    },
    {
      "entropy": 0.914535841345787,
      "epoch": 4.821477241755247,
      "grad_norm": 0.10512598603963852,
      "learning_rate": 8.117683897994704e-05,
      "loss": 1.345,
      "mean_token_accuracy": 0.7527642354369164,
      "num_tokens": 1271443060.0,
      "step": 17690
    },
    {
      "entropy": 0.9209541931748391,
      "epoch": 4.82420278004906,
      "grad_norm": 0.1139163002371788,
      "learning_rate": 8.115334504573101e-05,
      "loss": 1.3513,
      "mean_token_accuracy": 0.751273301243782,
      "num_tokens": 1272171307.0,
      "step": 17700
    },
    {
      "entropy": 0.9186633139848709,
      "epoch": 4.826928318342873,
      "grad_norm": 0.09583355486392975,
      "learning_rate": 8.112983986330491e-05,
      "loss": 1.3585,
      "mean_token_accuracy": 0.7448722943663597,
      "num_tokens": 1272881173.0,
      "step": 17710
    },
    {
      "entropy": 0.9142263114452363,
      "epoch": 4.829653856636686,
      "grad_norm": 0.10262507200241089,
      "learning_rate": 8.11063234411555e-05,
      "loss": 1.3448,
      "mean_token_accuracy": 0.7498973265290261,
      "num_tokens": 1273592477.0,
      "step": 17720
    },
    {
      "entropy": 0.9207109525799752,
      "epoch": 4.8323793949304985,
      "grad_norm": 0.11521895229816437,
      "learning_rate": 8.10827957877736e-05,
      "loss": 1.3563,
      "mean_token_accuracy": 0.75123840123415,
      "num_tokens": 1274318734.0,
      "step": 17730
    },
    {
      "entropy": 0.9218509048223495,
      "epoch": 4.8351049332243115,
      "grad_norm": 0.13229934871196747,
      "learning_rate": 8.105925691165413e-05,
      "loss": 1.3546,
      "mean_token_accuracy": 0.7465604782104492,
      "num_tokens": 1275026553.0,
      "step": 17740
    },
    {
      "entropy": 0.9194739490747452,
      "epoch": 4.837830471518124,
      "grad_norm": 0.10044396668672562,
      "learning_rate": 8.103570682129599e-05,
      "loss": 1.3457,
      "mean_token_accuracy": 0.7495238274335861,
      "num_tokens": 1275753007.0,
      "step": 17750
    },
    {
      "entropy": 0.9178076699376106,
      "epoch": 4.840556009811938,
      "grad_norm": 0.19709278643131256,
      "learning_rate": 8.101214552520216e-05,
      "loss": 1.3497,
      "mean_token_accuracy": 0.7485316529870033,
      "num_tokens": 1276479010.0,
      "step": 17760
    },
    {
      "entropy": 0.9250484704971313,
      "epoch": 4.843281548105751,
      "grad_norm": 0.10807700455188751,
      "learning_rate": 8.09885730318797e-05,
      "loss": 1.3558,
      "mean_token_accuracy": 0.7500222280621529,
      "num_tokens": 1277216478.0,
      "step": 17770
    },
    {
      "entropy": 0.9244115814566612,
      "epoch": 4.846007086399564,
      "grad_norm": 0.11148900538682938,
      "learning_rate": 8.096498934983965e-05,
      "loss": 1.357,
      "mean_token_accuracy": 0.7489386513829231,
      "num_tokens": 1277935610.0,
      "step": 17780
    },
    {
      "entropy": 0.916721199452877,
      "epoch": 4.848732624693377,
      "grad_norm": 0.11641812324523926,
      "learning_rate": 8.094139448759714e-05,
      "loss": 1.3495,
      "mean_token_accuracy": 0.7474601671099663,
      "num_tokens": 1278649268.0,
      "step": 17790
    },
    {
      "entropy": 0.9256363198161125,
      "epoch": 4.85145816298719,
      "grad_norm": 0.10508560389280319,
      "learning_rate": 8.09177884536713e-05,
      "loss": 1.3573,
      "mean_token_accuracy": 0.7465876549482345,
      "num_tokens": 1279367778.0,
      "step": 17800
    },
    {
      "entropy": 0.923135606944561,
      "epoch": 4.854183701281003,
      "grad_norm": 0.10860153287649155,
      "learning_rate": 8.089417125658534e-05,
      "loss": 1.3525,
      "mean_token_accuracy": 0.7471994385123253,
      "num_tokens": 1280084792.0,
      "step": 17810
    },
    {
      "entropy": 0.9226791754364967,
      "epoch": 4.856909239574816,
      "grad_norm": 0.13328388333320618,
      "learning_rate": 8.087054290486647e-05,
      "loss": 1.3472,
      "mean_token_accuracy": 0.7541562110185623,
      "num_tokens": 1280822314.0,
      "step": 17820
    },
    {
      "entropy": 0.918493464589119,
      "epoch": 4.859634777868629,
      "grad_norm": 0.10711206495761871,
      "learning_rate": 8.08469034070459e-05,
      "loss": 1.354,
      "mean_token_accuracy": 0.7474129155278206,
      "num_tokens": 1281535740.0,
      "step": 17830
    },
    {
      "entropy": 0.9164285495877266,
      "epoch": 4.862360316162442,
      "grad_norm": 0.10846658051013947,
      "learning_rate": 8.082325277165889e-05,
      "loss": 1.3501,
      "mean_token_accuracy": 0.7508197382092476,
      "num_tokens": 1282261076.0,
      "step": 17840
    },
    {
      "entropy": 0.9170690849423409,
      "epoch": 4.865085854456255,
      "grad_norm": 0.09901037067174911,
      "learning_rate": 8.079959100724477e-05,
      "loss": 1.3603,
      "mean_token_accuracy": 0.7489598989486694,
      "num_tokens": 1282972515.0,
      "step": 17850
    },
    {
      "entropy": 0.917238111793995,
      "epoch": 4.867811392750069,
      "grad_norm": 0.09574832022190094,
      "learning_rate": 8.077591812234684e-05,
      "loss": 1.3527,
      "mean_token_accuracy": 0.7521451890468598,
      "num_tokens": 1283694638.0,
      "step": 17860
    },
    {
      "entropy": 0.9155911713838577,
      "epoch": 4.8705369310438815,
      "grad_norm": 0.10780677199363708,
      "learning_rate": 8.075223412551237e-05,
      "loss": 1.3377,
      "mean_token_accuracy": 0.747342175245285,
      "num_tokens": 1284407095.0,
      "step": 17870
    },
    {
      "entropy": 0.9142779275774956,
      "epoch": 4.8732624693376945,
      "grad_norm": 0.10568727552890778,
      "learning_rate": 8.072853902529273e-05,
      "loss": 1.3432,
      "mean_token_accuracy": 0.7461394801735878,
      "num_tokens": 1285111455.0,
      "step": 17880
    },
    {
      "entropy": 0.9211603671312332,
      "epoch": 4.875988007631507,
      "grad_norm": 0.10505899041891098,
      "learning_rate": 8.070483283024326e-05,
      "loss": 1.3515,
      "mean_token_accuracy": 0.7452557429671287,
      "num_tokens": 1285819802.0,
      "step": 17890
    },
    {
      "entropy": 0.9161028012633323,
      "epoch": 4.87871354592532,
      "grad_norm": 0.09776589274406433,
      "learning_rate": 8.068111554892329e-05,
      "loss": 1.3378,
      "mean_token_accuracy": 0.7473795488476753,
      "num_tokens": 1286540207.0,
      "step": 17900
    },
    {
      "entropy": 0.9268019363284111,
      "epoch": 4.881439084219133,
      "grad_norm": 0.10650679469108582,
      "learning_rate": 8.065738718989618e-05,
      "loss": 1.3563,
      "mean_token_accuracy": 0.7462029054760932,
      "num_tokens": 1287262389.0,
      "step": 17910
    },
    {
      "entropy": 0.9144074723124505,
      "epoch": 4.884164622512946,
      "grad_norm": 0.11430129408836365,
      "learning_rate": 8.063364776172928e-05,
      "loss": 1.3382,
      "mean_token_accuracy": 0.7461496248841286,
      "num_tokens": 1287964225.0,
      "step": 17920
    },
    {
      "entropy": 0.9132382705807686,
      "epoch": 4.886890160806759,
      "grad_norm": 0.10328994691371918,
      "learning_rate": 8.060989727299392e-05,
      "loss": 1.3437,
      "mean_token_accuracy": 0.7519752576947212,
      "num_tokens": 1288697565.0,
      "step": 17930
    },
    {
      "entropy": 0.9186933055520058,
      "epoch": 4.889615699100572,
      "grad_norm": 0.09449103474617004,
      "learning_rate": 8.058613573226545e-05,
      "loss": 1.3518,
      "mean_token_accuracy": 0.7483037784695625,
      "num_tokens": 1289420743.0,
      "step": 17940
    },
    {
      "entropy": 0.9115293383598327,
      "epoch": 4.892341237394385,
      "grad_norm": 0.0965585708618164,
      "learning_rate": 8.05623631481232e-05,
      "loss": 1.3454,
      "mean_token_accuracy": 0.7499361336231232,
      "num_tokens": 1290150062.0,
      "step": 17950
    },
    {
      "entropy": 0.9113005891442298,
      "epoch": 4.895066775688198,
      "grad_norm": 0.10072828084230423,
      "learning_rate": 8.053857952915049e-05,
      "loss": 1.3366,
      "mean_token_accuracy": 0.7560128375887871,
      "num_tokens": 1290885707.0,
      "step": 17960
    },
    {
      "entropy": 0.9238735303282738,
      "epoch": 4.897792313982011,
      "grad_norm": 0.1140623539686203,
      "learning_rate": 8.051478488393463e-05,
      "loss": 1.3482,
      "mean_token_accuracy": 0.7446290031075478,
      "num_tokens": 1291589482.0,
      "step": 17970
    },
    {
      "entropy": 0.919769535958767,
      "epoch": 4.900517852275825,
      "grad_norm": 0.1134912520647049,
      "learning_rate": 8.049097922106687e-05,
      "loss": 1.3518,
      "mean_token_accuracy": 0.7496406942605972,
      "num_tokens": 1292313088.0,
      "step": 17980
    },
    {
      "entropy": 0.9129861995577813,
      "epoch": 4.903243390569638,
      "grad_norm": 0.11114895343780518,
      "learning_rate": 8.046716254914248e-05,
      "loss": 1.3437,
      "mean_token_accuracy": 0.7519585803151131,
      "num_tokens": 1293035418.0,
      "step": 17990
    },
    {
      "entropy": 0.9183290138840675,
      "epoch": 4.905968928863451,
      "grad_norm": 0.1106867641210556,
      "learning_rate": 8.044333487676072e-05,
      "loss": 1.3452,
      "mean_token_accuracy": 0.7493987768888474,
      "num_tokens": 1293751640.0,
      "step": 18000
    },
    {
      "entropy": 0.918357215821743,
      "epoch": 4.908694467157264,
      "grad_norm": 0.10817532241344452,
      "learning_rate": 8.041949621252478e-05,
      "loss": 1.3452,
      "mean_token_accuracy": 0.745792293548584,
      "num_tokens": 1294454616.0,
      "step": 18010
    },
    {
      "entropy": 0.911582262814045,
      "epoch": 4.911420005451077,
      "grad_norm": 0.1894681602716446,
      "learning_rate": 8.039564656504181e-05,
      "loss": 1.3403,
      "mean_token_accuracy": 0.753081364929676,
      "num_tokens": 1295178063.0,
      "step": 18020
    },
    {
      "entropy": 0.9186591416597366,
      "epoch": 4.9141455437448895,
      "grad_norm": 0.10621605813503265,
      "learning_rate": 8.0371785942923e-05,
      "loss": 1.355,
      "mean_token_accuracy": 0.7502840161323547,
      "num_tokens": 1295903866.0,
      "step": 18030
    },
    {
      "entropy": 0.9134357139468193,
      "epoch": 4.9168710820387025,
      "grad_norm": 0.24842919409275055,
      "learning_rate": 8.034791435478341e-05,
      "loss": 1.3461,
      "mean_token_accuracy": 0.7519378915429116,
      "num_tokens": 1296626622.0,
      "step": 18040
    },
    {
      "entropy": 0.9197294548153877,
      "epoch": 4.919596620332515,
      "grad_norm": 0.11456424742937088,
      "learning_rate": 8.032403180924213e-05,
      "loss": 1.3572,
      "mean_token_accuracy": 0.7467811524868011,
      "num_tokens": 1297338065.0,
      "step": 18050
    },
    {
      "entropy": 0.9246993482112884,
      "epoch": 4.922322158626328,
      "grad_norm": 0.11779490858316422,
      "learning_rate": 8.030013831492218e-05,
      "loss": 1.3583,
      "mean_token_accuracy": 0.7469172552227974,
      "num_tokens": 1298050724.0,
      "step": 18060
    },
    {
      "entropy": 0.9166697457432746,
      "epoch": 4.925047696920142,
      "grad_norm": 0.1063537448644638,
      "learning_rate": 8.027623388045052e-05,
      "loss": 1.3419,
      "mean_token_accuracy": 0.7515771239995956,
      "num_tokens": 1298779329.0,
      "step": 18070
    },
    {
      "entropy": 0.9245874509215355,
      "epoch": 4.927773235213955,
      "grad_norm": 0.10378915816545486,
      "learning_rate": 8.025231851445807e-05,
      "loss": 1.3544,
      "mean_token_accuracy": 0.7428726345300675,
      "num_tokens": 1299492520.0,
      "step": 18080
    },
    {
      "entropy": 0.9091032400727272,
      "epoch": 4.930498773507768,
      "grad_norm": 0.10003490000963211,
      "learning_rate": 8.02283922255797e-05,
      "loss": 1.3376,
      "mean_token_accuracy": 0.7529667928814888,
      "num_tokens": 1300227438.0,
      "step": 18090
    },
    {
      "entropy": 0.9176523268222809,
      "epoch": 4.933224311801581,
      "grad_norm": 0.12046157568693161,
      "learning_rate": 8.020445502245424e-05,
      "loss": 1.3508,
      "mean_token_accuracy": 0.7488412380218505,
      "num_tokens": 1300947339.0,
      "step": 18100
    },
    {
      "entropy": 0.9231064260005951,
      "epoch": 4.935949850095394,
      "grad_norm": 0.20513617992401123,
      "learning_rate": 8.018050691372441e-05,
      "loss": 1.3635,
      "mean_token_accuracy": 0.7462878093123436,
      "num_tokens": 1301649117.0,
      "step": 18110
    },
    {
      "entropy": 0.9169375166296959,
      "epoch": 4.938675388389207,
      "grad_norm": 0.12112665176391602,
      "learning_rate": 8.015654790803694e-05,
      "loss": 1.3543,
      "mean_token_accuracy": 0.7494909405708313,
      "num_tokens": 1302358306.0,
      "step": 18120
    },
    {
      "entropy": 0.9201806411147118,
      "epoch": 4.94140092668302,
      "grad_norm": 0.10498200356960297,
      "learning_rate": 8.013257801404244e-05,
      "loss": 1.3459,
      "mean_token_accuracy": 0.7431547686457634,
      "num_tokens": 1303063658.0,
      "step": 18130
    },
    {
      "entropy": 0.9263772219419479,
      "epoch": 4.944126464976833,
      "grad_norm": 0.10871928185224533,
      "learning_rate": 8.010859724039546e-05,
      "loss": 1.3562,
      "mean_token_accuracy": 0.7448802530765534,
      "num_tokens": 1303775553.0,
      "step": 18140
    },
    {
      "entropy": 0.9173216402530671,
      "epoch": 4.946852003270646,
      "grad_norm": 0.10729706287384033,
      "learning_rate": 8.008460559575448e-05,
      "loss": 1.3507,
      "mean_token_accuracy": 0.748846410214901,
      "num_tokens": 1304501939.0,
      "step": 18150
    },
    {
      "entropy": 0.9160275503993034,
      "epoch": 4.949577541564459,
      "grad_norm": 0.10040333122015,
      "learning_rate": 8.006060308878193e-05,
      "loss": 1.3359,
      "mean_token_accuracy": 0.7510499969124794,
      "num_tokens": 1305233162.0,
      "step": 18160
    },
    {
      "entropy": 0.9199049517512321,
      "epoch": 4.952303079858272,
      "grad_norm": 0.10592690855264664,
      "learning_rate": 8.003658972814414e-05,
      "loss": 1.3551,
      "mean_token_accuracy": 0.7488234281539917,
      "num_tokens": 1305949328.0,
      "step": 18170
    },
    {
      "entropy": 0.9183659970760345,
      "epoch": 4.955028618152085,
      "grad_norm": 0.12657269835472107,
      "learning_rate": 8.001256552251135e-05,
      "loss": 1.3443,
      "mean_token_accuracy": 0.7549128860235215,
      "num_tokens": 1306693379.0,
      "step": 18180
    },
    {
      "entropy": 0.9196929454803466,
      "epoch": 4.9577541564458985,
      "grad_norm": 0.10157456994056702,
      "learning_rate": 7.998853048055772e-05,
      "loss": 1.3466,
      "mean_token_accuracy": 0.7506722420454025,
      "num_tokens": 1307417569.0,
      "step": 18190
    },
    {
      "entropy": 0.9178652420639992,
      "epoch": 4.960479694739711,
      "grad_norm": 0.10534054785966873,
      "learning_rate": 7.996448461096135e-05,
      "loss": 1.3502,
      "mean_token_accuracy": 0.7453674659132957,
      "num_tokens": 1308127786.0,
      "step": 18200
    },
    {
      "entropy": 0.925149618089199,
      "epoch": 4.963205233033524,
      "grad_norm": 0.10649655014276505,
      "learning_rate": 7.994042792240421e-05,
      "loss": 1.3593,
      "mean_token_accuracy": 0.7453730344772339,
      "num_tokens": 1308844093.0,
      "step": 18210
    },
    {
      "entropy": 0.9176445603370667,
      "epoch": 4.965930771327337,
      "grad_norm": 0.11317556351423264,
      "learning_rate": 7.991636042357224e-05,
      "loss": 1.3465,
      "mean_token_accuracy": 0.7458223670721054,
      "num_tokens": 1309565166.0,
      "step": 18220
    },
    {
      "entropy": 0.9187493264675141,
      "epoch": 4.96865630962115,
      "grad_norm": 0.13759465515613556,
      "learning_rate": 7.989228212315516e-05,
      "loss": 1.352,
      "mean_token_accuracy": 0.7450497880578041,
      "num_tokens": 1310275785.0,
      "step": 18230
    },
    {
      "entropy": 0.9150622397661209,
      "epoch": 4.971381847914963,
      "grad_norm": 0.110161691904068,
      "learning_rate": 7.986819302984673e-05,
      "loss": 1.3584,
      "mean_token_accuracy": 0.74679996073246,
      "num_tokens": 1310983551.0,
      "step": 18240
    },
    {
      "entropy": 0.9159223407506942,
      "epoch": 4.974107386208776,
      "grad_norm": 0.10151788592338562,
      "learning_rate": 7.984409315234454e-05,
      "loss": 1.3443,
      "mean_token_accuracy": 0.7486445784568787,
      "num_tokens": 1311706970.0,
      "step": 18250
    },
    {
      "entropy": 0.9219341740012169,
      "epoch": 4.976832924502589,
      "grad_norm": 0.09841329604387283,
      "learning_rate": 7.981998249935004e-05,
      "loss": 1.3534,
      "mean_token_accuracy": 0.7474974930286408,
      "num_tokens": 1312424580.0,
      "step": 18260
    },
    {
      "entropy": 0.9184188976883888,
      "epoch": 4.979558462796402,
      "grad_norm": 0.10171353816986084,
      "learning_rate": 7.979586107956864e-05,
      "loss": 1.3523,
      "mean_token_accuracy": 0.7439409390091896,
      "num_tokens": 1313129730.0,
      "step": 18270
    },
    {
      "entropy": 0.9176041468977928,
      "epoch": 4.982284001090215,
      "grad_norm": 0.1016509085893631,
      "learning_rate": 7.977172890170961e-05,
      "loss": 1.3523,
      "mean_token_accuracy": 0.7485899671912193,
      "num_tokens": 1313844412.0,
      "step": 18280
    },
    {
      "entropy": 0.9199514031410218,
      "epoch": 4.985009539384029,
      "grad_norm": 0.10359106212854385,
      "learning_rate": 7.974758597448608e-05,
      "loss": 1.3427,
      "mean_token_accuracy": 0.7483617037534713,
      "num_tokens": 1314567086.0,
      "step": 18290
    },
    {
      "entropy": 0.9154374852776528,
      "epoch": 4.987735077677842,
      "grad_norm": 0.10706379264593124,
      "learning_rate": 7.97234323066151e-05,
      "loss": 1.346,
      "mean_token_accuracy": 0.7493002146482468,
      "num_tokens": 1315289883.0,
      "step": 18300
    },
    {
      "entropy": 0.9149588719010353,
      "epoch": 4.990460615971655,
      "grad_norm": 0.11768409609794617,
      "learning_rate": 7.969926790681759e-05,
      "loss": 1.3384,
      "mean_token_accuracy": 0.750345665216446,
      "num_tokens": 1316008435.0,
      "step": 18310
    },
    {
      "entropy": 0.9264756590127945,
      "epoch": 4.993186154265468,
      "grad_norm": 0.10559073090553284,
      "learning_rate": 7.96750927838183e-05,
      "loss": 1.3621,
      "mean_token_accuracy": 0.7422044023871421,
      "num_tokens": 1316722462.0,
      "step": 18320
    },
    {
      "entropy": 0.9273223444819451,
      "epoch": 4.995911692559281,
      "grad_norm": 0.10167667269706726,
      "learning_rate": 7.965090694634592e-05,
      "loss": 1.3559,
      "mean_token_accuracy": 0.7435720786452293,
      "num_tokens": 1317433726.0,
      "step": 18330
    },
    {
      "entropy": 0.9213701769709587,
      "epoch": 4.9986372308530935,
      "grad_norm": 0.094398133456707,
      "learning_rate": 7.962671040313297e-05,
      "loss": 1.3506,
      "mean_token_accuracy": 0.7460046291351319,
      "num_tokens": 1318147407.0,
      "step": 18340
    },
    {
      "entropy": 0.9180674225091934,
      "epoch": 5.0013627691469065,
      "grad_norm": 0.1045336127281189,
      "learning_rate": 7.960250316291582e-05,
      "loss": 1.3457,
      "mean_token_accuracy": 0.7494631335139275,
      "num_tokens": 1318866389.0,
      "step": 18350
    },
    {
      "entropy": 0.9117056980729104,
      "epoch": 5.004088307440719,
      "grad_norm": 0.10809372365474701,
      "learning_rate": 7.957828523443473e-05,
      "loss": 1.3441,
      "mean_token_accuracy": 0.7481998920440673,
      "num_tokens": 1319571770.0,
      "step": 18360
    },
    {
      "entropy": 0.9222692921757698,
      "epoch": 5.006813845734532,
      "grad_norm": 0.1019110381603241,
      "learning_rate": 7.955405662643384e-05,
      "loss": 1.3476,
      "mean_token_accuracy": 0.7470973148941994,
      "num_tokens": 1320297544.0,
      "step": 18370
    },
    {
      "entropy": 0.9072718366980552,
      "epoch": 5.009539384028345,
      "grad_norm": 0.11654374748468399,
      "learning_rate": 7.952981734766107e-05,
      "loss": 1.3293,
      "mean_token_accuracy": 0.7586778029799461,
      "num_tokens": 1321046846.0,
      "step": 18380
    },
    {
      "entropy": 0.9028063580393791,
      "epoch": 5.012264922322158,
      "grad_norm": 0.10506423562765121,
      "learning_rate": 7.95055674068683e-05,
      "loss": 1.3258,
      "mean_token_accuracy": 0.751988410949707,
      "num_tokens": 1321773389.0,
      "step": 18390
    },
    {
      "entropy": 0.910191185772419,
      "epoch": 5.014990460615972,
      "grad_norm": 0.12120980024337769,
      "learning_rate": 7.948130681281115e-05,
      "loss": 1.3397,
      "mean_token_accuracy": 0.7533672720193862,
      "num_tokens": 1322502982.0,
      "step": 18400
    },
    {
      "entropy": 0.9012519985437393,
      "epoch": 5.017715998909785,
      "grad_norm": 0.11411497741937637,
      "learning_rate": 7.945703557424916e-05,
      "loss": 1.3365,
      "mean_token_accuracy": 0.7553670346736908,
      "num_tokens": 1323232428.0,
      "step": 18410
    },
    {
      "entropy": 0.9155339032411576,
      "epoch": 5.020441537203598,
      "grad_norm": 0.10370244085788727,
      "learning_rate": 7.94327536999457e-05,
      "loss": 1.3406,
      "mean_token_accuracy": 0.7517688021063804,
      "num_tokens": 1323954276.0,
      "step": 18420
    },
    {
      "entropy": 0.9078729420900344,
      "epoch": 5.023167075497411,
      "grad_norm": 0.12481620907783508,
      "learning_rate": 7.940846119866794e-05,
      "loss": 1.3351,
      "mean_token_accuracy": 0.751006792485714,
      "num_tokens": 1324674401.0,
      "step": 18430
    },
    {
      "entropy": 0.9182976961135865,
      "epoch": 5.025892613791224,
      "grad_norm": 0.12291675806045532,
      "learning_rate": 7.938415807918693e-05,
      "loss": 1.3463,
      "mean_token_accuracy": 0.7464212417602539,
      "num_tokens": 1325386716.0,
      "step": 18440
    },
    {
      "entropy": 0.9183823958039283,
      "epoch": 5.028618152085037,
      "grad_norm": 0.1038193628191948,
      "learning_rate": 7.935984435027756e-05,
      "loss": 1.3419,
      "mean_token_accuracy": 0.7476585775613784,
      "num_tokens": 1326108125.0,
      "step": 18450
    },
    {
      "entropy": 0.909471395611763,
      "epoch": 5.03134369037885,
      "grad_norm": 0.10644439607858658,
      "learning_rate": 7.933552002071851e-05,
      "loss": 1.3351,
      "mean_token_accuracy": 0.750180697441101,
      "num_tokens": 1326826661.0,
      "step": 18460
    },
    {
      "entropy": 0.9158590584993362,
      "epoch": 5.034069228672663,
      "grad_norm": 0.11362117528915405,
      "learning_rate": 7.931118509929231e-05,
      "loss": 1.3389,
      "mean_token_accuracy": 0.7460008159279823,
      "num_tokens": 1327538266.0,
      "step": 18470
    },
    {
      "entropy": 0.8958799347281456,
      "epoch": 5.036794766966476,
      "grad_norm": 0.1139308363199234,
      "learning_rate": 7.928683959478535e-05,
      "loss": 1.3218,
      "mean_token_accuracy": 0.7529631853103638,
      "num_tokens": 1328256807.0,
      "step": 18480
    },
    {
      "entropy": 0.9072618633508682,
      "epoch": 5.039520305260289,
      "grad_norm": 0.11525868624448776,
      "learning_rate": 7.926248351598776e-05,
      "loss": 1.3404,
      "mean_token_accuracy": 0.7525811344385147,
      "num_tokens": 1328975615.0,
      "step": 18490
    },
    {
      "entropy": 0.9116022393107415,
      "epoch": 5.0422458435541015,
      "grad_norm": 0.10684821754693985,
      "learning_rate": 7.923811687169355e-05,
      "loss": 1.3379,
      "mean_token_accuracy": 0.7463834941387176,
      "num_tokens": 1329682884.0,
      "step": 18500
    },
    {
      "entropy": 0.9065963506698609,
      "epoch": 5.044971381847915,
      "grad_norm": 0.1166912093758583,
      "learning_rate": 7.921373967070052e-05,
      "loss": 1.33,
      "mean_token_accuracy": 0.7524808749556542,
      "num_tokens": 1330409855.0,
      "step": 18510
    },
    {
      "entropy": 0.9096634238958359,
      "epoch": 5.047696920141728,
      "grad_norm": 0.11363758146762848,
      "learning_rate": 7.91893519218103e-05,
      "loss": 1.3413,
      "mean_token_accuracy": 0.7469534277915955,
      "num_tokens": 1331119708.0,
      "step": 18520
    },
    {
      "entropy": 0.9015256181359291,
      "epoch": 5.050422458435541,
      "grad_norm": 0.11088868230581284,
      "learning_rate": 7.91649536338283e-05,
      "loss": 1.3305,
      "mean_token_accuracy": 0.7513253837823868,
      "num_tokens": 1331842581.0,
      "step": 18530
    },
    {
      "entropy": 0.9056093961000442,
      "epoch": 5.053147996729354,
      "grad_norm": 0.10830790549516678,
      "learning_rate": 7.914054481556379e-05,
      "loss": 1.3384,
      "mean_token_accuracy": 0.751427860558033,
      "num_tokens": 1332555238.0,
      "step": 18540
    },
    {
      "entropy": 0.8952996760606766,
      "epoch": 5.055873535023167,
      "grad_norm": 0.11326909065246582,
      "learning_rate": 7.911612547582977e-05,
      "loss": 1.3195,
      "mean_token_accuracy": 0.7594034299254417,
      "num_tokens": 1333296230.0,
      "step": 18550
    },
    {
      "entropy": 0.9200844272971154,
      "epoch": 5.05859907331698,
      "grad_norm": 0.11373654752969742,
      "learning_rate": 7.909169562344308e-05,
      "loss": 1.3413,
      "mean_token_accuracy": 0.7461709216237068,
      "num_tokens": 1334003940.0,
      "step": 18560
    },
    {
      "entropy": 0.9090452209115029,
      "epoch": 5.061324611610793,
      "grad_norm": 0.12129314243793488,
      "learning_rate": 7.906725526722436e-05,
      "loss": 1.3422,
      "mean_token_accuracy": 0.750136086344719,
      "num_tokens": 1334713399.0,
      "step": 18570
    },
    {
      "entropy": 0.9078034043312073,
      "epoch": 5.064050149904606,
      "grad_norm": 0.10491450875997543,
      "learning_rate": 7.904280441599801e-05,
      "loss": 1.3326,
      "mean_token_accuracy": 0.7513612136244774,
      "num_tokens": 1335432967.0,
      "step": 18580
    },
    {
      "entropy": 0.8964256376028061,
      "epoch": 5.066775688198419,
      "grad_norm": 0.12562423944473267,
      "learning_rate": 7.901834307859228e-05,
      "loss": 1.3241,
      "mean_token_accuracy": 0.7554881185293197,
      "num_tokens": 1336158037.0,
      "step": 18590
    },
    {
      "entropy": 0.911538377404213,
      "epoch": 5.069501226492232,
      "grad_norm": 0.11411541700363159,
      "learning_rate": 7.899387126383913e-05,
      "loss": 1.3378,
      "mean_token_accuracy": 0.7462952226400376,
      "num_tokens": 1336867959.0,
      "step": 18600
    },
    {
      "entropy": 0.9089008331298828,
      "epoch": 5.072226764786045,
      "grad_norm": 0.12004344910383224,
      "learning_rate": 7.896938898057435e-05,
      "loss": 1.3371,
      "mean_token_accuracy": 0.7475906565785408,
      "num_tokens": 1337572400.0,
      "step": 18610
    },
    {
      "entropy": 0.9116697743535042,
      "epoch": 5.074952303079859,
      "grad_norm": 0.12836812436580658,
      "learning_rate": 7.894489623763753e-05,
      "loss": 1.3405,
      "mean_token_accuracy": 0.7458344906568527,
      "num_tokens": 1338281950.0,
      "step": 18620
    },
    {
      "entropy": 0.9137950226664543,
      "epoch": 5.077677841373672,
      "grad_norm": 0.13870224356651306,
      "learning_rate": 7.892039304387195e-05,
      "loss": 1.338,
      "mean_token_accuracy": 0.749837937951088,
      "num_tokens": 1339003188.0,
      "step": 18630
    },
    {
      "entropy": 0.9065858721733093,
      "epoch": 5.080403379667485,
      "grad_norm": 0.149013489484787,
      "learning_rate": 7.889587940812477e-05,
      "loss": 1.3362,
      "mean_token_accuracy": 0.7499426275491714,
      "num_tokens": 1339715479.0,
      "step": 18640
    },
    {
      "entropy": 0.9157001972198486,
      "epoch": 5.0831289179612975,
      "grad_norm": 0.2263483703136444,
      "learning_rate": 7.887135533924683e-05,
      "loss": 1.3412,
      "mean_token_accuracy": 0.7490920692682266,
      "num_tokens": 1340428244.0,
      "step": 18650
    },
    {
      "entropy": 0.9115339174866677,
      "epoch": 5.0858544562551105,
      "grad_norm": 0.1228538453578949,
      "learning_rate": 7.884682084609282e-05,
      "loss": 1.3426,
      "mean_token_accuracy": 0.744776040315628,
      "num_tokens": 1341133482.0,
      "step": 18660
    },
    {
      "entropy": 0.9042779088020325,
      "epoch": 5.088579994548923,
      "grad_norm": 0.11652397364377975,
      "learning_rate": 7.882227593752112e-05,
      "loss": 1.3357,
      "mean_token_accuracy": 0.7493279471993446,
      "num_tokens": 1341846045.0,
      "step": 18670
    },
    {
      "entropy": 0.9101129621267319,
      "epoch": 5.091305532842736,
      "grad_norm": 0.11319631338119507,
      "learning_rate": 7.879772062239392e-05,
      "loss": 1.3366,
      "mean_token_accuracy": 0.7525549560785294,
      "num_tokens": 1342571187.0,
      "step": 18680
    },
    {
      "entropy": 0.90872992426157,
      "epoch": 5.094031071136549,
      "grad_norm": 0.11274012178182602,
      "learning_rate": 7.877315490957714e-05,
      "loss": 1.3427,
      "mean_token_accuracy": 0.7483867600560188,
      "num_tokens": 1343278501.0,
      "step": 18690
    },
    {
      "entropy": 0.9083260744810104,
      "epoch": 5.096756609430362,
      "grad_norm": 0.1183522492647171,
      "learning_rate": 7.874857880794046e-05,
      "loss": 1.3385,
      "mean_token_accuracy": 0.7478100255131721,
      "num_tokens": 1343993725.0,
      "step": 18700
    },
    {
      "entropy": 0.9029194608330726,
      "epoch": 5.099482147724175,
      "grad_norm": 0.11869072169065475,
      "learning_rate": 7.872399232635731e-05,
      "loss": 1.3314,
      "mean_token_accuracy": 0.7562534168362618,
      "num_tokens": 1344730125.0,
      "step": 18710
    },
    {
      "entropy": 0.9123732477426529,
      "epoch": 5.102207686017988,
      "grad_norm": 0.11277887225151062,
      "learning_rate": 7.869939547370489e-05,
      "loss": 1.3319,
      "mean_token_accuracy": 0.7512045651674271,
      "num_tokens": 1345458796.0,
      "step": 18720
    },
    {
      "entropy": 0.9051063165068627,
      "epoch": 5.104933224311802,
      "grad_norm": 0.11701159179210663,
      "learning_rate": 7.86747882588641e-05,
      "loss": 1.3336,
      "mean_token_accuracy": 0.750642716884613,
      "num_tokens": 1346171647.0,
      "step": 18730
    },
    {
      "entropy": 0.9102186009287834,
      "epoch": 5.107658762605615,
      "grad_norm": 0.10872722417116165,
      "learning_rate": 7.865017069071964e-05,
      "loss": 1.3351,
      "mean_token_accuracy": 0.7451282486319541,
      "num_tokens": 1346885311.0,
      "step": 18740
    },
    {
      "entropy": 0.9089977607131005,
      "epoch": 5.110384300899428,
      "grad_norm": 0.22637620568275452,
      "learning_rate": 7.86255427781599e-05,
      "loss": 1.341,
      "mean_token_accuracy": 0.7501268580555915,
      "num_tokens": 1347610287.0,
      "step": 18750
    },
    {
      "entropy": 0.9072558596730232,
      "epoch": 5.113109839193241,
      "grad_norm": 0.12592224776744843,
      "learning_rate": 7.8600904530077e-05,
      "loss": 1.3343,
      "mean_token_accuracy": 0.7507593974471092,
      "num_tokens": 1348329760.0,
      "step": 18760
    },
    {
      "entropy": 0.9108141466975213,
      "epoch": 5.115835377487054,
      "grad_norm": 0.1176564022898674,
      "learning_rate": 7.857625595536685e-05,
      "loss": 1.3366,
      "mean_token_accuracy": 0.7493161335587502,
      "num_tokens": 1349043593.0,
      "step": 18770
    },
    {
      "entropy": 0.907894492149353,
      "epoch": 5.118560915780867,
      "grad_norm": 0.11292760074138641,
      "learning_rate": 7.855159706292903e-05,
      "loss": 1.3312,
      "mean_token_accuracy": 0.7481828153133392,
      "num_tokens": 1349756028.0,
      "step": 18780
    },
    {
      "entropy": 0.9055081680417061,
      "epoch": 5.12128645407468,
      "grad_norm": 0.10565032064914703,
      "learning_rate": 7.852692786166688e-05,
      "loss": 1.3384,
      "mean_token_accuracy": 0.753260900080204,
      "num_tokens": 1350472891.0,
      "step": 18790
    },
    {
      "entropy": 0.899952945113182,
      "epoch": 5.124011992368493,
      "grad_norm": 0.1228262260556221,
      "learning_rate": 7.850224836048742e-05,
      "loss": 1.3287,
      "mean_token_accuracy": 0.7540143489837646,
      "num_tokens": 1351189567.0,
      "step": 18800
    },
    {
      "entropy": 0.9097656950354576,
      "epoch": 5.1267375306623055,
      "grad_norm": 0.11419453471899033,
      "learning_rate": 7.847755856830146e-05,
      "loss": 1.3434,
      "mean_token_accuracy": 0.7453598141670227,
      "num_tokens": 1351886225.0,
      "step": 18810
    },
    {
      "entropy": 0.9096739515662193,
      "epoch": 5.1294630689561185,
      "grad_norm": 0.11032206565141678,
      "learning_rate": 7.845285849402344e-05,
      "loss": 1.3323,
      "mean_token_accuracy": 0.7500687360763549,
      "num_tokens": 1352610800.0,
      "step": 18820
    },
    {
      "entropy": 0.9082689002156258,
      "epoch": 5.132188607249931,
      "grad_norm": 0.12134139239788055,
      "learning_rate": 7.842814814657158e-05,
      "loss": 1.3315,
      "mean_token_accuracy": 0.7526320442557335,
      "num_tokens": 1353335250.0,
      "step": 18830
    },
    {
      "entropy": 0.9058627471327781,
      "epoch": 5.134914145543745,
      "grad_norm": 0.12179207056760788,
      "learning_rate": 7.840342753486778e-05,
      "loss": 1.3355,
      "mean_token_accuracy": 0.7513252019882202,
      "num_tokens": 1354051481.0,
      "step": 18840
    },
    {
      "entropy": 0.9101318687200546,
      "epoch": 5.137639683837558,
      "grad_norm": 0.10935749858617783,
      "learning_rate": 7.837869666783766e-05,
      "loss": 1.3398,
      "mean_token_accuracy": 0.7486051961779594,
      "num_tokens": 1354754677.0,
      "step": 18850
    },
    {
      "entropy": 0.9036993905901909,
      "epoch": 5.140365222131371,
      "grad_norm": 0.11682439595460892,
      "learning_rate": 7.835395555441051e-05,
      "loss": 1.3364,
      "mean_token_accuracy": 0.7515853494405746,
      "num_tokens": 1355465373.0,
      "step": 18860
    },
    {
      "entropy": 0.8975047558546067,
      "epoch": 5.143090760425184,
      "grad_norm": 0.11143290996551514,
      "learning_rate": 7.832920420351937e-05,
      "loss": 1.3299,
      "mean_token_accuracy": 0.7522084057331085,
      "num_tokens": 1356175972.0,
      "step": 18870
    },
    {
      "entropy": 0.9004489123821259,
      "epoch": 5.145816298718997,
      "grad_norm": 0.10541433840990067,
      "learning_rate": 7.830444262410093e-05,
      "loss": 1.3291,
      "mean_token_accuracy": 0.7541014179587364,
      "num_tokens": 1356898950.0,
      "step": 18880
    },
    {
      "entropy": 0.9021969020366669,
      "epoch": 5.14854183701281,
      "grad_norm": 0.10437975078821182,
      "learning_rate": 7.82796708250956e-05,
      "loss": 1.3341,
      "mean_token_accuracy": 0.7506950855255127,
      "num_tokens": 1357601726.0,
      "step": 18890
    },
    {
      "entropy": 0.912436631321907,
      "epoch": 5.151267375306623,
      "grad_norm": 0.1150716245174408,
      "learning_rate": 7.825488881544748e-05,
      "loss": 1.3452,
      "mean_token_accuracy": 0.7441218435764313,
      "num_tokens": 1358305410.0,
      "step": 18900
    },
    {
      "entropy": 0.9159964561462403,
      "epoch": 5.153992913600436,
      "grad_norm": 0.12094953656196594,
      "learning_rate": 7.823009660410434e-05,
      "loss": 1.347,
      "mean_token_accuracy": 0.7476379483938217,
      "num_tokens": 1359013224.0,
      "step": 18910
    },
    {
      "entropy": 0.9071477174758911,
      "epoch": 5.156718451894249,
      "grad_norm": 0.11563536524772644,
      "learning_rate": 7.820529420001763e-05,
      "loss": 1.3357,
      "mean_token_accuracy": 0.7493469908833503,
      "num_tokens": 1359721073.0,
      "step": 18920
    },
    {
      "entropy": 0.9029750868678093,
      "epoch": 5.159443990188062,
      "grad_norm": 0.11280180513858795,
      "learning_rate": 7.818048161214253e-05,
      "loss": 1.3255,
      "mean_token_accuracy": 0.7580265924334526,
      "num_tokens": 1360460944.0,
      "step": 18930
    },
    {
      "entropy": 0.9054033905267715,
      "epoch": 5.162169528481876,
      "grad_norm": 0.11178267002105713,
      "learning_rate": 7.815565884943782e-05,
      "loss": 1.3372,
      "mean_token_accuracy": 0.750056803226471,
      "num_tokens": 1361166446.0,
      "step": 18940
    },
    {
      "entropy": 0.9099515557289124,
      "epoch": 5.1648950667756885,
      "grad_norm": 0.10570986568927765,
      "learning_rate": 7.8130825920866e-05,
      "loss": 1.3349,
      "mean_token_accuracy": 0.7498923137784004,
      "num_tokens": 1361890170.0,
      "step": 18950
    },
    {
      "entropy": 0.9100850179791451,
      "epoch": 5.1676206050695015,
      "grad_norm": 0.12696321308612823,
      "learning_rate": 7.810598283539328e-05,
      "loss": 1.3329,
      "mean_token_accuracy": 0.7483788222074509,
      "num_tokens": 1362603199.0,
      "step": 18960
    },
    {
      "entropy": 0.9101712808012963,
      "epoch": 5.170346143363314,
      "grad_norm": 0.11161109805107117,
      "learning_rate": 7.808112960198942e-05,
      "loss": 1.3422,
      "mean_token_accuracy": 0.7523481532931328,
      "num_tokens": 1363323415.0,
      "step": 18970
    },
    {
      "entropy": 0.9075639858841896,
      "epoch": 5.173071681657127,
      "grad_norm": 0.12539154291152954,
      "learning_rate": 7.805626622962796e-05,
      "loss": 1.3276,
      "mean_token_accuracy": 0.750854168832302,
      "num_tokens": 1364041718.0,
      "step": 18980
    },
    {
      "entropy": 0.9018291965126991,
      "epoch": 5.17579721995094,
      "grad_norm": 0.12013975530862808,
      "learning_rate": 7.803139272728606e-05,
      "loss": 1.3307,
      "mean_token_accuracy": 0.7495744436979294,
      "num_tokens": 1364746616.0,
      "step": 18990
    },
    {
      "entropy": 0.9024452075362206,
      "epoch": 5.178522758244753,
      "grad_norm": 0.10841753333806992,
      "learning_rate": 7.800650910394449e-05,
      "loss": 1.3216,
      "mean_token_accuracy": 0.7567202374339104,
      "num_tokens": 1365485662.0,
      "step": 19000
    },
    {
      "entropy": 0.9038829073309899,
      "epoch": 5.181248296538566,
      "grad_norm": 0.1268831491470337,
      "learning_rate": 7.798161536858776e-05,
      "loss": 1.3349,
      "mean_token_accuracy": 0.7507214352488518,
      "num_tokens": 1366196113.0,
      "step": 19010
    },
    {
      "entropy": 0.9127470672130584,
      "epoch": 5.183973834832379,
      "grad_norm": 0.13608817756175995,
      "learning_rate": 7.795671153020399e-05,
      "loss": 1.3392,
      "mean_token_accuracy": 0.747468601167202,
      "num_tokens": 1366909657.0,
      "step": 19020
    },
    {
      "entropy": 0.9076567769050599,
      "epoch": 5.186699373126192,
      "grad_norm": 0.12499534338712692,
      "learning_rate": 7.793179759778493e-05,
      "loss": 1.3357,
      "mean_token_accuracy": 0.7521659284830093,
      "num_tokens": 1367635085.0,
      "step": 19030
    },
    {
      "entropy": 0.9101962640881538,
      "epoch": 5.189424911420005,
      "grad_norm": 0.11534728854894638,
      "learning_rate": 7.790687358032598e-05,
      "loss": 1.3409,
      "mean_token_accuracy": 0.7470502510666848,
      "num_tokens": 1368341214.0,
      "step": 19040
    },
    {
      "entropy": 0.9091815829277039,
      "epoch": 5.192150449713819,
      "grad_norm": 0.1163393184542656,
      "learning_rate": 7.788193948682621e-05,
      "loss": 1.3431,
      "mean_token_accuracy": 0.7499541327357292,
      "num_tokens": 1369055070.0,
      "step": 19050
    },
    {
      "entropy": 0.9086731627583504,
      "epoch": 5.194875988007632,
      "grad_norm": 0.11986309289932251,
      "learning_rate": 7.78569953262883e-05,
      "loss": 1.3369,
      "mean_token_accuracy": 0.748068630695343,
      "num_tokens": 1369771537.0,
      "step": 19060
    },
    {
      "entropy": 0.9071047112345696,
      "epoch": 5.197601526301445,
      "grad_norm": 0.11011476814746857,
      "learning_rate": 7.78320411077186e-05,
      "loss": 1.3312,
      "mean_token_accuracy": 0.751863069832325,
      "num_tokens": 1370488216.0,
      "step": 19070
    },
    {
      "entropy": 0.9116240307688713,
      "epoch": 5.200327064595258,
      "grad_norm": 0.11632122099399567,
      "learning_rate": 7.780707684012703e-05,
      "loss": 1.3452,
      "mean_token_accuracy": 0.7497341588139534,
      "num_tokens": 1371201818.0,
      "step": 19080
    },
    {
      "entropy": 0.9063777029514313,
      "epoch": 5.203052602889071,
      "grad_norm": 0.12040931731462479,
      "learning_rate": 7.77821025325272e-05,
      "loss": 1.3354,
      "mean_token_accuracy": 0.7512107893824578,
      "num_tokens": 1371923557.0,
      "step": 19090
    },
    {
      "entropy": 0.9083833083510399,
      "epoch": 5.205778141182884,
      "grad_norm": 0.12313888221979141,
      "learning_rate": 7.775711819393632e-05,
      "loss": 1.332,
      "mean_token_accuracy": 0.7498325526714325,
      "num_tokens": 1372652851.0,
      "step": 19100
    },
    {
      "entropy": 0.909971933066845,
      "epoch": 5.208503679476697,
      "grad_norm": 0.11367330700159073,
      "learning_rate": 7.77321238333752e-05,
      "loss": 1.3374,
      "mean_token_accuracy": 0.7517801567912101,
      "num_tokens": 1373386848.0,
      "step": 19110
    },
    {
      "entropy": 0.906111478805542,
      "epoch": 5.2112292177705095,
      "grad_norm": 0.12350302189588547,
      "learning_rate": 7.770711945986833e-05,
      "loss": 1.3326,
      "mean_token_accuracy": 0.7504727497696877,
      "num_tokens": 1374109890.0,
      "step": 19120
    },
    {
      "entropy": 0.8956575006246567,
      "epoch": 5.2139547560643225,
      "grad_norm": 0.11455515772104263,
      "learning_rate": 7.768210508244373e-05,
      "loss": 1.3251,
      "mean_token_accuracy": 0.7561988785862923,
      "num_tokens": 1374830846.0,
      "step": 19130
    },
    {
      "entropy": 0.9134457930922508,
      "epoch": 5.216680294358135,
      "grad_norm": 0.11794273555278778,
      "learning_rate": 7.765708071013311e-05,
      "loss": 1.3429,
      "mean_token_accuracy": 0.7462093397974968,
      "num_tokens": 1375540653.0,
      "step": 19140
    },
    {
      "entropy": 0.9088801175355912,
      "epoch": 5.219405832651948,
      "grad_norm": 0.127328559756279,
      "learning_rate": 7.763204635197174e-05,
      "loss": 1.3346,
      "mean_token_accuracy": 0.7491170018911362,
      "num_tokens": 1376262815.0,
      "step": 19150
    },
    {
      "entropy": 0.9111427307128906,
      "epoch": 5.222131370945762,
      "grad_norm": 0.11400771141052246,
      "learning_rate": 7.760700201699854e-05,
      "loss": 1.3447,
      "mean_token_accuracy": 0.7470154613256454,
      "num_tokens": 1376970324.0,
      "step": 19160
    },
    {
      "entropy": 0.911037626862526,
      "epoch": 5.224856909239575,
      "grad_norm": 0.1092766523361206,
      "learning_rate": 7.758194771425594e-05,
      "loss": 1.3417,
      "mean_token_accuracy": 0.7472645550966263,
      "num_tokens": 1377674775.0,
      "step": 19170
    },
    {
      "entropy": 0.9015596866607666,
      "epoch": 5.227582447533388,
      "grad_norm": 0.12264671176671982,
      "learning_rate": 7.755688345279009e-05,
      "loss": 1.3255,
      "mean_token_accuracy": 0.7519594937562942,
      "num_tokens": 1378389712.0,
      "step": 19180
    },
    {
      "entropy": 0.9012925669550895,
      "epoch": 5.230307985827201,
      "grad_norm": 0.11462859064340591,
      "learning_rate": 7.753180924165069e-05,
      "loss": 1.3331,
      "mean_token_accuracy": 0.7517343029379845,
      "num_tokens": 1379112505.0,
      "step": 19190
    },
    {
      "entropy": 0.910705478489399,
      "epoch": 5.233033524121014,
      "grad_norm": 0.12070734798908234,
      "learning_rate": 7.750672508989096e-05,
      "loss": 1.3299,
      "mean_token_accuracy": 0.7470805883407593,
      "num_tokens": 1379823856.0,
      "step": 19200
    },
    {
      "entropy": 0.9054174587130547,
      "epoch": 5.235759062414827,
      "grad_norm": 0.10295546054840088,
      "learning_rate": 7.748163100656782e-05,
      "loss": 1.3338,
      "mean_token_accuracy": 0.7534317210316658,
      "num_tokens": 1380557353.0,
      "step": 19210
    },
    {
      "entropy": 0.9041800126433372,
      "epoch": 5.23848460070864,
      "grad_norm": 0.11778882890939713,
      "learning_rate": 7.745652700074172e-05,
      "loss": 1.3328,
      "mean_token_accuracy": 0.7520045042037964,
      "num_tokens": 1381270670.0,
      "step": 19220
    },
    {
      "entropy": 0.8997292220592499,
      "epoch": 5.241210139002453,
      "grad_norm": 0.11303021013736725,
      "learning_rate": 7.743141308147668e-05,
      "loss": 1.3269,
      "mean_token_accuracy": 0.7499917581677437,
      "num_tokens": 1381972787.0,
      "step": 19230
    },
    {
      "entropy": 0.9059125691652298,
      "epoch": 5.243935677296266,
      "grad_norm": 0.11100378632545471,
      "learning_rate": 7.740628925784032e-05,
      "loss": 1.3365,
      "mean_token_accuracy": 0.7547996297478676,
      "num_tokens": 1382716022.0,
      "step": 19240
    },
    {
      "entropy": 0.9083168834447861,
      "epoch": 5.246661215590079,
      "grad_norm": 0.10881959646940231,
      "learning_rate": 7.738115553890387e-05,
      "loss": 1.3371,
      "mean_token_accuracy": 0.7476097568869591,
      "num_tokens": 1383423565.0,
      "step": 19250
    },
    {
      "entropy": 0.9086048856377602,
      "epoch": 5.249386753883892,
      "grad_norm": 0.11474138498306274,
      "learning_rate": 7.735601193374205e-05,
      "loss": 1.3357,
      "mean_token_accuracy": 0.7469335675239563,
      "num_tokens": 1384123967.0,
      "step": 19260
    },
    {
      "entropy": 0.9057020366191864,
      "epoch": 5.2521122921777055,
      "grad_norm": 0.13058039546012878,
      "learning_rate": 7.733085845143322e-05,
      "loss": 1.3322,
      "mean_token_accuracy": 0.751021246612072,
      "num_tokens": 1384851520.0,
      "step": 19270
    },
    {
      "entropy": 0.9090562105178833,
      "epoch": 5.254837830471518,
      "grad_norm": 0.12455740571022034,
      "learning_rate": 7.730569510105929e-05,
      "loss": 1.3444,
      "mean_token_accuracy": 0.7508804053068161,
      "num_tokens": 1385575365.0,
      "step": 19280
    },
    {
      "entropy": 0.9198357537388802,
      "epoch": 5.257563368765331,
      "grad_norm": 0.11248283833265305,
      "learning_rate": 7.72805218917057e-05,
      "loss": 1.3463,
      "mean_token_accuracy": 0.7451340645551682,
      "num_tokens": 1386289570.0,
      "step": 19290
    },
    {
      "entropy": 0.9151126220822334,
      "epoch": 5.260288907059144,
      "grad_norm": 0.11552846431732178,
      "learning_rate": 7.72553388324615e-05,
      "loss": 1.3381,
      "mean_token_accuracy": 0.7482043281197548,
      "num_tokens": 1387004623.0,
      "step": 19300
    },
    {
      "entropy": 0.9030871108174324,
      "epoch": 5.263014445352957,
      "grad_norm": 0.123524971306324,
      "learning_rate": 7.723014593241926e-05,
      "loss": 1.3348,
      "mean_token_accuracy": 0.7522234320640564,
      "num_tokens": 1387729084.0,
      "step": 19310
    },
    {
      "entropy": 0.899794852733612,
      "epoch": 5.26573998364677,
      "grad_norm": 0.11385549604892731,
      "learning_rate": 7.720494320067513e-05,
      "loss": 1.3272,
      "mean_token_accuracy": 0.7524638131260872,
      "num_tokens": 1388438859.0,
      "step": 19320
    },
    {
      "entropy": 0.9128438696265221,
      "epoch": 5.268465521940583,
      "grad_norm": 0.10310929268598557,
      "learning_rate": 7.717973064632875e-05,
      "loss": 1.3402,
      "mean_token_accuracy": 0.7502741917967797,
      "num_tokens": 1389167010.0,
      "step": 19330
    },
    {
      "entropy": 0.9036715939640999,
      "epoch": 5.271191060234396,
      "grad_norm": 0.11425105482339859,
      "learning_rate": 7.715450827848342e-05,
      "loss": 1.3275,
      "mean_token_accuracy": 0.7534997895359993,
      "num_tokens": 1389901009.0,
      "step": 19340
    },
    {
      "entropy": 0.9070032447576523,
      "epoch": 5.273916598528209,
      "grad_norm": 0.11248745769262314,
      "learning_rate": 7.712927610624586e-05,
      "loss": 1.3385,
      "mean_token_accuracy": 0.7467876300215721,
      "num_tokens": 1390611861.0,
      "step": 19350
    },
    {
      "entropy": 0.9052289351820946,
      "epoch": 5.276642136822022,
      "grad_norm": 0.11316084861755371,
      "learning_rate": 7.71040341387264e-05,
      "loss": 1.3306,
      "mean_token_accuracy": 0.7510544061660767,
      "num_tokens": 1391337025.0,
      "step": 19360
    },
    {
      "entropy": 0.9025282666087151,
      "epoch": 5.279367675115836,
      "grad_norm": 0.11133657395839691,
      "learning_rate": 7.707878238503891e-05,
      "loss": 1.3274,
      "mean_token_accuracy": 0.7547363251447677,
      "num_tokens": 1392063071.0,
      "step": 19370
    },
    {
      "entropy": 0.9065290451049804,
      "epoch": 5.282093213409649,
      "grad_norm": 0.11806612461805344,
      "learning_rate": 7.705352085430075e-05,
      "loss": 1.3335,
      "mean_token_accuracy": 0.7478181257843971,
      "num_tokens": 1392775824.0,
      "step": 19380
    },
    {
      "entropy": 0.9054453298449516,
      "epoch": 5.284818751703462,
      "grad_norm": 0.11038026958703995,
      "learning_rate": 7.702824955563284e-05,
      "loss": 1.334,
      "mean_token_accuracy": 0.7522917613387108,
      "num_tokens": 1393496991.0,
      "step": 19390
    },
    {
      "entropy": 0.9151970610022545,
      "epoch": 5.287544289997275,
      "grad_norm": 0.14004695415496826,
      "learning_rate": 7.700296849815963e-05,
      "loss": 1.3487,
      "mean_token_accuracy": 0.7442385584115982,
      "num_tokens": 1394207687.0,
      "step": 19400
    },
    {
      "entropy": 0.9074890226125717,
      "epoch": 5.290269828291088,
      "grad_norm": 0.11384852975606918,
      "learning_rate": 7.697767769100908e-05,
      "loss": 1.3296,
      "mean_token_accuracy": 0.7558194622397423,
      "num_tokens": 1394955201.0,
      "step": 19410
    },
    {
      "entropy": 0.9050102978944778,
      "epoch": 5.2929953665849006,
      "grad_norm": 0.12064824998378754,
      "learning_rate": 7.695237714331266e-05,
      "loss": 1.3359,
      "mean_token_accuracy": 0.7510082855820656,
      "num_tokens": 1395673016.0,
      "step": 19420
    },
    {
      "entropy": 0.9079841196537017,
      "epoch": 5.2957209048787135,
      "grad_norm": 0.1407109946012497,
      "learning_rate": 7.692706686420543e-05,
      "loss": 1.3343,
      "mean_token_accuracy": 0.7475971832871438,
      "num_tokens": 1396382824.0,
      "step": 19430
    },
    {
      "entropy": 0.9041830107569695,
      "epoch": 5.298446443172526,
      "grad_norm": 0.11657635122537613,
      "learning_rate": 7.690174686282583e-05,
      "loss": 1.336,
      "mean_token_accuracy": 0.7495165050029755,
      "num_tokens": 1397096108.0,
      "step": 19440
    },
    {
      "entropy": 0.9078656613826752,
      "epoch": 5.301171981466339,
      "grad_norm": 0.11287423968315125,
      "learning_rate": 7.687641714831593e-05,
      "loss": 1.3352,
      "mean_token_accuracy": 0.7502091526985168,
      "num_tokens": 1397816095.0,
      "step": 19450
    },
    {
      "entropy": 0.9061649590730667,
      "epoch": 5.303897519760152,
      "grad_norm": 0.12110287696123123,
      "learning_rate": 7.685107772982125e-05,
      "loss": 1.3337,
      "mean_token_accuracy": 0.7503016650676727,
      "num_tokens": 1398529791.0,
      "step": 19460
    },
    {
      "entropy": 0.9036917462944984,
      "epoch": 5.306623058053965,
      "grad_norm": 0.11222954094409943,
      "learning_rate": 7.682572861649083e-05,
      "loss": 1.334,
      "mean_token_accuracy": 0.7517042055726051,
      "num_tokens": 1399256259.0,
      "step": 19470
    },
    {
      "entropy": 0.9025445148348809,
      "epoch": 5.309348596347778,
      "grad_norm": 0.11223101615905762,
      "learning_rate": 7.68003698174772e-05,
      "loss": 1.3321,
      "mean_token_accuracy": 0.7515106126666069,
      "num_tokens": 1399978047.0,
      "step": 19480
    },
    {
      "entropy": 0.9033101335167885,
      "epoch": 5.312074134641592,
      "grad_norm": 0.1127723678946495,
      "learning_rate": 7.677500134193643e-05,
      "loss": 1.3277,
      "mean_token_accuracy": 0.7530519783496856,
      "num_tokens": 1400699195.0,
      "step": 19490
    },
    {
      "entropy": 0.904048889875412,
      "epoch": 5.314799672935405,
      "grad_norm": 0.10669442266225815,
      "learning_rate": 7.674962319902799e-05,
      "loss": 1.3292,
      "mean_token_accuracy": 0.7526533991098404,
      "num_tokens": 1401423595.0,
      "step": 19500
    },
    {
      "entropy": 0.905977538228035,
      "epoch": 5.317525211229218,
      "grad_norm": 0.11804743111133575,
      "learning_rate": 7.672423539791494e-05,
      "loss": 1.3295,
      "mean_token_accuracy": 0.7526730731129646,
      "num_tokens": 1402157190.0,
      "step": 19510
    },
    {
      "entropy": 0.895224104821682,
      "epoch": 5.320250749523031,
      "grad_norm": 0.12127908319234848,
      "learning_rate": 7.669883794776376e-05,
      "loss": 1.3305,
      "mean_token_accuracy": 0.7575635626912117,
      "num_tokens": 1402892572.0,
      "step": 19520
    },
    {
      "entropy": 0.9105286493897438,
      "epoch": 5.322976287816844,
      "grad_norm": 0.1054353341460228,
      "learning_rate": 7.667343085774448e-05,
      "loss": 1.3381,
      "mean_token_accuracy": 0.7508665218949317,
      "num_tokens": 1403611219.0,
      "step": 19530
    },
    {
      "entropy": 0.9054419875144959,
      "epoch": 5.325701826110657,
      "grad_norm": 0.11713731288909912,
      "learning_rate": 7.664801413703052e-05,
      "loss": 1.339,
      "mean_token_accuracy": 0.7495303899049759,
      "num_tokens": 1404317519.0,
      "step": 19540
    },
    {
      "entropy": 0.9082285046577454,
      "epoch": 5.32842736440447,
      "grad_norm": 0.12444167584180832,
      "learning_rate": 7.662258779479888e-05,
      "loss": 1.3311,
      "mean_token_accuracy": 0.7541111633181572,
      "num_tokens": 1405040140.0,
      "step": 19550
    },
    {
      "entropy": 0.8985153838992119,
      "epoch": 5.331152902698283,
      "grad_norm": 0.12595868110656738,
      "learning_rate": 7.659715184022994e-05,
      "loss": 1.3269,
      "mean_token_accuracy": 0.7527410835027695,
      "num_tokens": 1405749398.0,
      "step": 19560
    },
    {
      "entropy": 0.9099251523613929,
      "epoch": 5.333878440992096,
      "grad_norm": 0.1393590271472931,
      "learning_rate": 7.657170628250759e-05,
      "loss": 1.3372,
      "mean_token_accuracy": 0.7476893916726113,
      "num_tokens": 1406467895.0,
      "step": 19570
    },
    {
      "entropy": 0.9051065519452095,
      "epoch": 5.336603979285909,
      "grad_norm": 0.1308859884738922,
      "learning_rate": 7.654625113081922e-05,
      "loss": 1.3348,
      "mean_token_accuracy": 0.7513506069779397,
      "num_tokens": 1407192334.0,
      "step": 19580
    },
    {
      "entropy": 0.9081906601786613,
      "epoch": 5.339329517579722,
      "grad_norm": 0.11979203671216965,
      "learning_rate": 7.652078639435563e-05,
      "loss": 1.3389,
      "mean_token_accuracy": 0.751957580447197,
      "num_tokens": 1407909539.0,
      "step": 19590
    },
    {
      "entropy": 0.9147018074989319,
      "epoch": 5.342055055873535,
      "grad_norm": 0.11424284428358078,
      "learning_rate": 7.649531208231112e-05,
      "loss": 1.3482,
      "mean_token_accuracy": 0.7451479762792588,
      "num_tokens": 1408621393.0,
      "step": 19600
    },
    {
      "entropy": 0.9017274841666222,
      "epoch": 5.344780594167348,
      "grad_norm": 0.11710764467716217,
      "learning_rate": 7.646982820388342e-05,
      "loss": 1.3283,
      "mean_token_accuracy": 0.7538257449865341,
      "num_tokens": 1409343631.0,
      "step": 19610
    },
    {
      "entropy": 0.897904546558857,
      "epoch": 5.347506132461161,
      "grad_norm": 0.1083025112748146,
      "learning_rate": 7.644433476827373e-05,
      "loss": 1.325,
      "mean_token_accuracy": 0.7533201783895492,
      "num_tokens": 1410065428.0,
      "step": 19620
    },
    {
      "entropy": 0.8997597068548202,
      "epoch": 5.350231670754974,
      "grad_norm": 0.11617015302181244,
      "learning_rate": 7.64188317846867e-05,
      "loss": 1.3257,
      "mean_token_accuracy": 0.7566874668002128,
      "num_tokens": 1410795942.0,
      "step": 19630
    },
    {
      "entropy": 0.905792772769928,
      "epoch": 5.352957209048787,
      "grad_norm": 0.11008379608392715,
      "learning_rate": 7.639331926233042e-05,
      "loss": 1.3321,
      "mean_token_accuracy": 0.753586845099926,
      "num_tokens": 1411517166.0,
      "step": 19640
    },
    {
      "entropy": 0.9120403468608856,
      "epoch": 5.3556827473426,
      "grad_norm": 0.11209972947835922,
      "learning_rate": 7.636779721041642e-05,
      "loss": 1.3424,
      "mean_token_accuracy": 0.7472355544567109,
      "num_tokens": 1412235008.0,
      "step": 19650
    },
    {
      "entropy": 0.9027108520269393,
      "epoch": 5.358408285636413,
      "grad_norm": 0.1391863375902176,
      "learning_rate": 7.63422656381597e-05,
      "loss": 1.3313,
      "mean_token_accuracy": 0.75594282746315,
      "num_tokens": 1412962667.0,
      "step": 19660
    },
    {
      "entropy": 0.9091307118535041,
      "epoch": 5.361133823930226,
      "grad_norm": 0.12177189439535141,
      "learning_rate": 7.631672455477865e-05,
      "loss": 1.3396,
      "mean_token_accuracy": 0.7468419194221496,
      "num_tokens": 1413675189.0,
      "step": 19670
    },
    {
      "entropy": 0.9070631563663483,
      "epoch": 5.363859362224039,
      "grad_norm": 0.11187789589166641,
      "learning_rate": 7.629117396949512e-05,
      "loss": 1.3434,
      "mean_token_accuracy": 0.7456631541252137,
      "num_tokens": 1414378068.0,
      "step": 19680
    },
    {
      "entropy": 0.900896580517292,
      "epoch": 5.366584900517852,
      "grad_norm": 0.1092860996723175,
      "learning_rate": 7.626561389153441e-05,
      "loss": 1.3325,
      "mean_token_accuracy": 0.7518083170056343,
      "num_tokens": 1415095124.0,
      "step": 19690
    },
    {
      "entropy": 0.9109518691897392,
      "epoch": 5.369310438811666,
      "grad_norm": 0.11021330207586288,
      "learning_rate": 7.624004433012524e-05,
      "loss": 1.3458,
      "mean_token_accuracy": 0.7492843508720398,
      "num_tokens": 1415809720.0,
      "step": 19700
    },
    {
      "entropy": 0.9164837971329689,
      "epoch": 5.372035977105479,
      "grad_norm": 0.11230812966823578,
      "learning_rate": 7.621446529449969e-05,
      "loss": 1.3441,
      "mean_token_accuracy": 0.7437252566218376,
      "num_tokens": 1416514761.0,
      "step": 19710
    },
    {
      "entropy": 0.9080846697092056,
      "epoch": 5.374761515399292,
      "grad_norm": 0.12813308835029602,
      "learning_rate": 7.618887679389337e-05,
      "loss": 1.3318,
      "mean_token_accuracy": 0.7514203503727913,
      "num_tokens": 1417231382.0,
      "step": 19720
    },
    {
      "entropy": 0.9045114368200302,
      "epoch": 5.3774870536931045,
      "grad_norm": 0.11684253811836243,
      "learning_rate": 7.616327883754521e-05,
      "loss": 1.3303,
      "mean_token_accuracy": 0.7518407315015793,
      "num_tokens": 1417952019.0,
      "step": 19730
    },
    {
      "entropy": 0.9064024522900581,
      "epoch": 5.3802125919869175,
      "grad_norm": 0.12381158024072647,
      "learning_rate": 7.613767143469762e-05,
      "loss": 1.3301,
      "mean_token_accuracy": 0.7515939101576805,
      "num_tokens": 1418679736.0,
      "step": 19740
    },
    {
      "entropy": 0.9053021788597106,
      "epoch": 5.38293813028073,
      "grad_norm": 0.11646660417318344,
      "learning_rate": 7.611205459459639e-05,
      "loss": 1.3325,
      "mean_token_accuracy": 0.7485539689660072,
      "num_tokens": 1419397767.0,
      "step": 19750
    },
    {
      "entropy": 0.9090729087591172,
      "epoch": 5.385663668574543,
      "grad_norm": 0.10986338555812836,
      "learning_rate": 7.608642832649069e-05,
      "loss": 1.3352,
      "mean_token_accuracy": 0.7499466896057129,
      "num_tokens": 1420111514.0,
      "step": 19760
    },
    {
      "entropy": 0.9054888308048248,
      "epoch": 5.388389206868356,
      "grad_norm": 0.11227500438690186,
      "learning_rate": 7.606079263963317e-05,
      "loss": 1.3394,
      "mean_token_accuracy": 0.7531944304704666,
      "num_tokens": 1420830287.0,
      "step": 19770
    },
    {
      "entropy": 0.9086398214101792,
      "epoch": 5.391114745162169,
      "grad_norm": 0.11145076900720596,
      "learning_rate": 7.603514754327983e-05,
      "loss": 1.3427,
      "mean_token_accuracy": 0.7472387075424194,
      "num_tokens": 1421546419.0,
      "step": 19780
    },
    {
      "entropy": 0.9121958896517753,
      "epoch": 5.393840283455982,
      "grad_norm": 0.11511708796024323,
      "learning_rate": 7.600949304669005e-05,
      "loss": 1.344,
      "mean_token_accuracy": 0.7537903472781181,
      "num_tokens": 1422276582.0,
      "step": 19790
    },
    {
      "entropy": 0.9117308780550957,
      "epoch": 5.396565821749796,
      "grad_norm": 0.10570119321346283,
      "learning_rate": 7.598382915912665e-05,
      "loss": 1.3452,
      "mean_token_accuracy": 0.7464302599430084,
      "num_tokens": 1422984287.0,
      "step": 19800
    },
    {
      "entropy": 0.8995277851819992,
      "epoch": 5.399291360043609,
      "grad_norm": 0.12177679687738419,
      "learning_rate": 7.595815588985581e-05,
      "loss": 1.3243,
      "mean_token_accuracy": 0.755612076818943,
      "num_tokens": 1423719401.0,
      "step": 19810
    },
    {
      "entropy": 0.9075158998370171,
      "epoch": 5.402016898337422,
      "grad_norm": 0.12687304615974426,
      "learning_rate": 7.593247324814712e-05,
      "loss": 1.3419,
      "mean_token_accuracy": 0.7462605714797974,
      "num_tokens": 1424415621.0,
      "step": 19820
    },
    {
      "entropy": 0.9051286518573761,
      "epoch": 5.404742436631235,
      "grad_norm": 0.11561757326126099,
      "learning_rate": 7.590678124327354e-05,
      "loss": 1.3298,
      "mean_token_accuracy": 0.7537098959088325,
      "num_tokens": 1425143985.0,
      "step": 19830
    },
    {
      "entropy": 0.9019662946462631,
      "epoch": 5.407467974925048,
      "grad_norm": 0.11109740287065506,
      "learning_rate": 7.58810798845114e-05,
      "loss": 1.3359,
      "mean_token_accuracy": 0.755416439473629,
      "num_tokens": 1425871385.0,
      "step": 19840
    },
    {
      "entropy": 0.9108938261866569,
      "epoch": 5.410193513218861,
      "grad_norm": 0.11039499938488007,
      "learning_rate": 7.585536918114041e-05,
      "loss": 1.3386,
      "mean_token_accuracy": 0.749568022787571,
      "num_tokens": 1426588716.0,
      "step": 19850
    },
    {
      "entropy": 0.8967614889144897,
      "epoch": 5.412919051512674,
      "grad_norm": 0.1311003565788269,
      "learning_rate": 7.582964914244368e-05,
      "loss": 1.3299,
      "mean_token_accuracy": 0.751678578555584,
      "num_tokens": 1427307684.0,
      "step": 19860
    },
    {
      "entropy": 0.9076550498604774,
      "epoch": 5.415644589806487,
      "grad_norm": 0.12214123457670212,
      "learning_rate": 7.580391977770767e-05,
      "loss": 1.337,
      "mean_token_accuracy": 0.7504000514745712,
      "num_tokens": 1428018267.0,
      "step": 19870
    },
    {
      "entropy": 0.9148858696222305,
      "epoch": 5.4183701281003,
      "grad_norm": 0.10823305696249008,
      "learning_rate": 7.577818109622221e-05,
      "loss": 1.3431,
      "mean_token_accuracy": 0.7481129378080368,
      "num_tokens": 1428738502.0,
      "step": 19880
    },
    {
      "entropy": 0.9061247304081916,
      "epoch": 5.4210956663941126,
      "grad_norm": 0.11874030530452728,
      "learning_rate": 7.575243310728048e-05,
      "loss": 1.33,
      "mean_token_accuracy": 0.7531487971544266,
      "num_tokens": 1429466540.0,
      "step": 19890
    },
    {
      "entropy": 0.9076259955763817,
      "epoch": 5.4238212046879255,
      "grad_norm": 0.12292955815792084,
      "learning_rate": 7.572667582017907e-05,
      "loss": 1.3379,
      "mean_token_accuracy": 0.7469023033976555,
      "num_tokens": 1430166146.0,
      "step": 19900
    },
    {
      "entropy": 0.909446969628334,
      "epoch": 5.4265467429817384,
      "grad_norm": 0.1153629794716835,
      "learning_rate": 7.570090924421786e-05,
      "loss": 1.3424,
      "mean_token_accuracy": 0.7452532842755317,
      "num_tokens": 1430864648.0,
      "step": 19910
    },
    {
      "entropy": 0.8984705403447151,
      "epoch": 5.429272281275552,
      "grad_norm": 0.10957809537649155,
      "learning_rate": 7.567513338870013e-05,
      "loss": 1.3293,
      "mean_token_accuracy": 0.7593380361795425,
      "num_tokens": 1431601094.0,
      "step": 19920
    },
    {
      "entropy": 0.9151330262422561,
      "epoch": 5.431997819569365,
      "grad_norm": 0.11466176062822342,
      "learning_rate": 7.564934826293251e-05,
      "loss": 1.3398,
      "mean_token_accuracy": 0.7458806976675987,
      "num_tokens": 1432311408.0,
      "step": 19930
    },
    {
      "entropy": 0.898513650894165,
      "epoch": 5.434723357863178,
      "grad_norm": 0.1184920072555542,
      "learning_rate": 7.562355387622493e-05,
      "loss": 1.3179,
      "mean_token_accuracy": 0.7580939635634423,
      "num_tokens": 1433044062.0,
      "step": 19940
    },
    {
      "entropy": 0.9102798148989677,
      "epoch": 5.437448896156991,
      "grad_norm": 0.11374345421791077,
      "learning_rate": 7.559775023789071e-05,
      "loss": 1.3422,
      "mean_token_accuracy": 0.7491901963949203,
      "num_tokens": 1433768093.0,
      "step": 19950
    },
    {
      "entropy": 0.9103214174509049,
      "epoch": 5.440174434450804,
      "grad_norm": 0.11783763766288757,
      "learning_rate": 7.557193735724651e-05,
      "loss": 1.3353,
      "mean_token_accuracy": 0.7503290638327599,
      "num_tokens": 1434487735.0,
      "step": 19960
    },
    {
      "entropy": 0.908466774225235,
      "epoch": 5.442899972744617,
      "grad_norm": 0.11903638392686844,
      "learning_rate": 7.55461152436123e-05,
      "loss": 1.3349,
      "mean_token_accuracy": 0.7494949519634246,
      "num_tokens": 1435203822.0,
      "step": 19970
    },
    {
      "entropy": 0.9091369539499283,
      "epoch": 5.44562551103843,
      "grad_norm": 0.11400353163480759,
      "learning_rate": 7.55202839063114e-05,
      "loss": 1.339,
      "mean_token_accuracy": 0.7500902906060218,
      "num_tokens": 1435922304.0,
      "step": 19980
    },
    {
      "entropy": 0.9097193807363511,
      "epoch": 5.448351049332243,
      "grad_norm": 0.11606765538454056,
      "learning_rate": 7.549444335467049e-05,
      "loss": 1.3421,
      "mean_token_accuracy": 0.7503934010863305,
      "num_tokens": 1436640597.0,
      "step": 19990
    },
    {
      "entropy": 0.9020377561450005,
      "epoch": 5.451076587626056,
      "grad_norm": 0.11023647338151932,
      "learning_rate": 7.54685935980195e-05,
      "loss": 1.3323,
      "mean_token_accuracy": 0.7534341529011727,
      "num_tokens": 1437353112.0,
      "step": 20000
    },
    {
      "entropy": 0.9162172496318817,
      "epoch": 5.453802125919869,
      "grad_norm": 0.11996352672576904,
      "learning_rate": 7.544273464569176e-05,
      "loss": 1.3452,
      "mean_token_accuracy": 0.7470343619585037,
      "num_tokens": 1438069947.0,
      "step": 20010
    },
    {
      "entropy": 0.9053322419524192,
      "epoch": 5.456527664213683,
      "grad_norm": 0.1083291694521904,
      "learning_rate": 7.541686650702387e-05,
      "loss": 1.3345,
      "mean_token_accuracy": 0.7503931105136872,
      "num_tokens": 1438792526.0,
      "step": 20020
    },
    {
      "entropy": 0.9153704836964607,
      "epoch": 5.459253202507496,
      "grad_norm": 0.11806406825780869,
      "learning_rate": 7.539098919135579e-05,
      "loss": 1.3447,
      "mean_token_accuracy": 0.7499242499470711,
      "num_tokens": 1439516853.0,
      "step": 20030
    },
    {
      "entropy": 0.8986057966947556,
      "epoch": 5.4619787408013085,
      "grad_norm": 0.12491049617528915,
      "learning_rate": 7.536510270803074e-05,
      "loss": 1.3206,
      "mean_token_accuracy": 0.7556749016046524,
      "num_tokens": 1440248947.0,
      "step": 20040
    },
    {
      "entropy": 0.9066046863794327,
      "epoch": 5.4647042790951215,
      "grad_norm": 0.11053416877985,
      "learning_rate": 7.533920706639531e-05,
      "loss": 1.3348,
      "mean_token_accuracy": 0.7530885234475135,
      "num_tokens": 1440979604.0,
      "step": 20050
    },
    {
      "entropy": 0.9050970584154129,
      "epoch": 5.467429817388934,
      "grad_norm": 0.12423629313707352,
      "learning_rate": 7.531330227579935e-05,
      "loss": 1.3317,
      "mean_token_accuracy": 0.7531575649976731,
      "num_tokens": 1441707103.0,
      "step": 20060
    },
    {
      "entropy": 0.907585859298706,
      "epoch": 5.470155355682747,
      "grad_norm": 0.12942558526992798,
      "learning_rate": 7.528738834559601e-05,
      "loss": 1.3357,
      "mean_token_accuracy": 0.7485477849841118,
      "num_tokens": 1442418576.0,
      "step": 20070
    },
    {
      "entropy": 0.907803401350975,
      "epoch": 5.47288089397656,
      "grad_norm": 0.11285819858312607,
      "learning_rate": 7.526146528514182e-05,
      "loss": 1.3434,
      "mean_token_accuracy": 0.7495760977268219,
      "num_tokens": 1443131298.0,
      "step": 20080
    },
    {
      "entropy": 0.9082187086343765,
      "epoch": 5.475606432270373,
      "grad_norm": 0.11220671236515045,
      "learning_rate": 7.523553310379648e-05,
      "loss": 1.3343,
      "mean_token_accuracy": 0.7479541718959808,
      "num_tokens": 1443845812.0,
      "step": 20090
    },
    {
      "entropy": 0.9146851256489754,
      "epoch": 5.478331970564186,
      "grad_norm": 0.12780430912971497,
      "learning_rate": 7.520959181092308e-05,
      "loss": 1.3447,
      "mean_token_accuracy": 0.7462508618831635,
      "num_tokens": 1444557250.0,
      "step": 20100
    },
    {
      "entropy": 0.9132054328918457,
      "epoch": 5.481057508857999,
      "grad_norm": 0.11686737835407257,
      "learning_rate": 7.518364141588797e-05,
      "loss": 1.3388,
      "mean_token_accuracy": 0.7521233156323432,
      "num_tokens": 1445282491.0,
      "step": 20110
    },
    {
      "entropy": 0.9029791802167892,
      "epoch": 5.483783047151812,
      "grad_norm": 0.11449231207370758,
      "learning_rate": 7.515768192806078e-05,
      "loss": 1.3311,
      "mean_token_accuracy": 0.7513266012072564,
      "num_tokens": 1446000947.0,
      "step": 20120
    },
    {
      "entropy": 0.9054128766059876,
      "epoch": 5.486508585445626,
      "grad_norm": 0.09959646314382553,
      "learning_rate": 7.513171335681442e-05,
      "loss": 1.3281,
      "mean_token_accuracy": 0.7520932570099831,
      "num_tokens": 1446727749.0,
      "step": 20130
    },
    {
      "entropy": 0.9029397517442703,
      "epoch": 5.489234123739439,
      "grad_norm": 0.13366763293743134,
      "learning_rate": 7.510573571152513e-05,
      "loss": 1.3295,
      "mean_token_accuracy": 0.7515622198581695,
      "num_tokens": 1447455668.0,
      "step": 20140
    },
    {
      "entropy": 0.9119272291660309,
      "epoch": 5.491959662033252,
      "grad_norm": 0.12279974669218063,
      "learning_rate": 7.507974900157233e-05,
      "loss": 1.3415,
      "mean_token_accuracy": 0.7486102536320687,
      "num_tokens": 1448171096.0,
      "step": 20150
    },
    {
      "entropy": 0.9111403718590736,
      "epoch": 5.494685200327065,
      "grad_norm": 0.11162228882312775,
      "learning_rate": 7.505375323633878e-05,
      "loss": 1.3483,
      "mean_token_accuracy": 0.7475664585828781,
      "num_tokens": 1448879233.0,
      "step": 20160
    },
    {
      "entropy": 0.8979356229305268,
      "epoch": 5.497410738620878,
      "grad_norm": 0.11792001128196716,
      "learning_rate": 7.502774842521054e-05,
      "loss": 1.3229,
      "mean_token_accuracy": 0.7544900998473167,
      "num_tokens": 1449597550.0,
      "step": 20170
    },
    {
      "entropy": 0.9082313403487206,
      "epoch": 5.500136276914691,
      "grad_norm": 0.10402727872133255,
      "learning_rate": 7.500173457757685e-05,
      "loss": 1.3398,
      "mean_token_accuracy": 0.7487385854125023,
      "num_tokens": 1450309282.0,
      "step": 20180
    },
    {
      "entropy": 0.9001877471804619,
      "epoch": 5.502861815208504,
      "grad_norm": 0.12189797312021255,
      "learning_rate": 7.497571170283026e-05,
      "loss": 1.3272,
      "mean_token_accuracy": 0.7528101876378059,
      "num_tokens": 1451019808.0,
      "step": 20190
    },
    {
      "entropy": 0.8964033558964729,
      "epoch": 5.5055873535023165,
      "grad_norm": 0.11383513361215591,
      "learning_rate": 7.49496798103666e-05,
      "loss": 1.3204,
      "mean_token_accuracy": 0.7546953797340393,
      "num_tokens": 1451744964.0,
      "step": 20200
    },
    {
      "entropy": 0.9120410367846489,
      "epoch": 5.5083128917961295,
      "grad_norm": 0.11335021257400513,
      "learning_rate": 7.492363890958489e-05,
      "loss": 1.3379,
      "mean_token_accuracy": 0.7478732347488404,
      "num_tokens": 1452457818.0,
      "step": 20210
    },
    {
      "entropy": 0.903009295463562,
      "epoch": 5.511038430089942,
      "grad_norm": 0.10975080728530884,
      "learning_rate": 7.489758900988748e-05,
      "loss": 1.3348,
      "mean_token_accuracy": 0.7547759979963302,
      "num_tokens": 1453177700.0,
      "step": 20220
    },
    {
      "entropy": 0.9056313022971153,
      "epoch": 5.513763968383756,
      "grad_norm": 0.12535880506038666,
      "learning_rate": 7.487153012067992e-05,
      "loss": 1.3367,
      "mean_token_accuracy": 0.7520108968019485,
      "num_tokens": 1453901511.0,
      "step": 20230
    },
    {
      "entropy": 0.9079207628965378,
      "epoch": 5.516489506677569,
      "grad_norm": 0.12012846022844315,
      "learning_rate": 7.484546225137103e-05,
      "loss": 1.3287,
      "mean_token_accuracy": 0.7541177779436111,
      "num_tokens": 1454632443.0,
      "step": 20240
    },
    {
      "entropy": 0.9155611962080001,
      "epoch": 5.519215044971382,
      "grad_norm": 0.11876795440912247,
      "learning_rate": 7.481938541137284e-05,
      "loss": 1.344,
      "mean_token_accuracy": 0.7458859100937844,
      "num_tokens": 1455349699.0,
      "step": 20250
    },
    {
      "entropy": 0.907274204492569,
      "epoch": 5.521940583265195,
      "grad_norm": 0.12015290558338165,
      "learning_rate": 7.479329961010065e-05,
      "loss": 1.3294,
      "mean_token_accuracy": 0.7453959077596665,
      "num_tokens": 1456063702.0,
      "step": 20260
    },
    {
      "entropy": 0.9086082488298416,
      "epoch": 5.524666121559008,
      "grad_norm": 0.11473998427391052,
      "learning_rate": 7.476720485697299e-05,
      "loss": 1.3398,
      "mean_token_accuracy": 0.7506502345204353,
      "num_tokens": 1456782137.0,
      "step": 20270
    },
    {
      "entropy": 0.9110724046826363,
      "epoch": 5.527391659852821,
      "grad_norm": 0.11411876976490021,
      "learning_rate": 7.47411011614116e-05,
      "loss": 1.3442,
      "mean_token_accuracy": 0.7489840254187584,
      "num_tokens": 1457505573.0,
      "step": 20280
    },
    {
      "entropy": 0.9120266690850258,
      "epoch": 5.530117198146634,
      "grad_norm": 0.12499787658452988,
      "learning_rate": 7.471498853284148e-05,
      "loss": 1.3405,
      "mean_token_accuracy": 0.7482805222272872,
      "num_tokens": 1458221670.0,
      "step": 20290
    },
    {
      "entropy": 0.9048104181885719,
      "epoch": 5.532842736440447,
      "grad_norm": 0.13739416003227234,
      "learning_rate": 7.468886698069085e-05,
      "loss": 1.3347,
      "mean_token_accuracy": 0.7490284994244576,
      "num_tokens": 1458939298.0,
      "step": 20300
    },
    {
      "entropy": 0.9044500857591629,
      "epoch": 5.53556827473426,
      "grad_norm": 0.1126444861292839,
      "learning_rate": 7.466273651439114e-05,
      "loss": 1.3326,
      "mean_token_accuracy": 0.7474282279610633,
      "num_tokens": 1459645367.0,
      "step": 20310
    },
    {
      "entropy": 0.9038434520363807,
      "epoch": 5.538293813028073,
      "grad_norm": 0.11752692610025406,
      "learning_rate": 7.463659714337699e-05,
      "loss": 1.3326,
      "mean_token_accuracy": 0.7536931246519089,
      "num_tokens": 1460374527.0,
      "step": 20320
    },
    {
      "entropy": 0.9099839940667153,
      "epoch": 5.541019351321886,
      "grad_norm": 0.11410974711179733,
      "learning_rate": 7.461044887708628e-05,
      "loss": 1.3342,
      "mean_token_accuracy": 0.751562948524952,
      "num_tokens": 1461096966.0,
      "step": 20330
    },
    {
      "entropy": 0.9014537543058395,
      "epoch": 5.543744889615699,
      "grad_norm": 0.11125433444976807,
      "learning_rate": 7.458429172496008e-05,
      "loss": 1.3297,
      "mean_token_accuracy": 0.7535984814167023,
      "num_tokens": 1461815635.0,
      "step": 20340
    },
    {
      "entropy": 0.9084295406937599,
      "epoch": 5.5464704279095125,
      "grad_norm": 0.11080422252416611,
      "learning_rate": 7.455812569644267e-05,
      "loss": 1.3341,
      "mean_token_accuracy": 0.7486643940210342,
      "num_tokens": 1462538299.0,
      "step": 20350
    },
    {
      "entropy": 0.9197900325059891,
      "epoch": 5.5491959662033254,
      "grad_norm": 0.11600863933563232,
      "learning_rate": 7.453195080098156e-05,
      "loss": 1.342,
      "mean_token_accuracy": 0.7477640584111214,
      "num_tokens": 1463266668.0,
      "step": 20360
    },
    {
      "entropy": 0.9118197932839394,
      "epoch": 5.551921504497138,
      "grad_norm": 0.1464001089334488,
      "learning_rate": 7.450576704802744e-05,
      "loss": 1.3389,
      "mean_token_accuracy": 0.7487740635871887,
      "num_tokens": 1463985261.0,
      "step": 20370
    },
    {
      "entropy": 0.9112050861120224,
      "epoch": 5.554647042790951,
      "grad_norm": 0.11449520289897919,
      "learning_rate": 7.44795744470342e-05,
      "loss": 1.3392,
      "mean_token_accuracy": 0.7453368082642555,
      "num_tokens": 1464694289.0,
      "step": 20380
    },
    {
      "entropy": 0.9063918381929398,
      "epoch": 5.557372581084764,
      "grad_norm": 0.10858428478240967,
      "learning_rate": 7.445337300745892e-05,
      "loss": 1.3307,
      "mean_token_accuracy": 0.7507871344685555,
      "num_tokens": 1465404383.0,
      "step": 20390
    },
    {
      "entropy": 0.9051426962018013,
      "epoch": 5.560098119378577,
      "grad_norm": 0.10692258179187775,
      "learning_rate": 7.442716273876188e-05,
      "loss": 1.3354,
      "mean_token_accuracy": 0.749012079834938,
      "num_tokens": 1466116990.0,
      "step": 20400
    },
    {
      "entropy": 0.9113408133387566,
      "epoch": 5.56282365767239,
      "grad_norm": 0.10736019909381866,
      "learning_rate": 7.440094365040656e-05,
      "loss": 1.3355,
      "mean_token_accuracy": 0.7532822921872139,
      "num_tokens": 1466842064.0,
      "step": 20410
    },
    {
      "entropy": 0.9093280717730522,
      "epoch": 5.565549195966203,
      "grad_norm": 0.11680605262517929,
      "learning_rate": 7.437471575185959e-05,
      "loss": 1.3378,
      "mean_token_accuracy": 0.747671315073967,
      "num_tokens": 1467556781.0,
      "step": 20420
    },
    {
      "entropy": 0.8958086788654327,
      "epoch": 5.568274734260016,
      "grad_norm": 0.12360627949237823,
      "learning_rate": 7.434847905259081e-05,
      "loss": 1.3228,
      "mean_token_accuracy": 0.7576392307877541,
      "num_tokens": 1468296360.0,
      "step": 20430
    },
    {
      "entropy": 0.902715815603733,
      "epoch": 5.571000272553829,
      "grad_norm": 0.11227795481681824,
      "learning_rate": 7.432223356207322e-05,
      "loss": 1.3327,
      "mean_token_accuracy": 0.7470630362629891,
      "num_tokens": 1469003601.0,
      "step": 20440
    },
    {
      "entropy": 0.8954124122858047,
      "epoch": 5.573725810847643,
      "grad_norm": 0.11043966561555862,
      "learning_rate": 7.429597928978302e-05,
      "loss": 1.327,
      "mean_token_accuracy": 0.7527671173214913,
      "num_tokens": 1469721159.0,
      "step": 20450
    },
    {
      "entropy": 0.8985204964876174,
      "epoch": 5.576451349141456,
      "grad_norm": 0.12268073856830597,
      "learning_rate": 7.426971624519953e-05,
      "loss": 1.3318,
      "mean_token_accuracy": 0.753266267478466,
      "num_tokens": 1470435623.0,
      "step": 20460
    },
    {
      "entropy": 0.9078370571136475,
      "epoch": 5.579176887435269,
      "grad_norm": 0.12063147127628326,
      "learning_rate": 7.424344443780534e-05,
      "loss": 1.335,
      "mean_token_accuracy": 0.7487000048160553,
      "num_tokens": 1471150721.0,
      "step": 20470
    },
    {
      "entropy": 0.9103540197014809,
      "epoch": 5.581902425729082,
      "grad_norm": 0.11219299584627151,
      "learning_rate": 7.421716387708605e-05,
      "loss": 1.338,
      "mean_token_accuracy": 0.7505401521921158,
      "num_tokens": 1471868754.0,
      "step": 20480
    },
    {
      "entropy": 0.9084721207618713,
      "epoch": 5.584627964022895,
      "grad_norm": 0.11898704618215561,
      "learning_rate": 7.419087457253057e-05,
      "loss": 1.3356,
      "mean_token_accuracy": 0.7494804993271827,
      "num_tokens": 1472592616.0,
      "step": 20490
    },
    {
      "entropy": 0.9035915836691857,
      "epoch": 5.587353502316708,
      "grad_norm": 0.12188675999641418,
      "learning_rate": 7.416457653363086e-05,
      "loss": 1.3249,
      "mean_token_accuracy": 0.7526019886136055,
      "num_tokens": 1473322061.0,
      "step": 20500
    },
    {
      "entropy": 0.9117862388491631,
      "epoch": 5.5900790406105205,
      "grad_norm": 0.1346140205860138,
      "learning_rate": 7.413826976988209e-05,
      "loss": 1.3362,
      "mean_token_accuracy": 0.747203154861927,
      "num_tokens": 1474026772.0,
      "step": 20510
    },
    {
      "entropy": 0.9080166593194008,
      "epoch": 5.5928045789043335,
      "grad_norm": 0.17116689682006836,
      "learning_rate": 7.411195429078258e-05,
      "loss": 1.3286,
      "mean_token_accuracy": 0.7503232747316361,
      "num_tokens": 1474752018.0,
      "step": 20520
    },
    {
      "entropy": 0.904227003455162,
      "epoch": 5.595530117198146,
      "grad_norm": 0.14302507042884827,
      "learning_rate": 7.408563010583377e-05,
      "loss": 1.3332,
      "mean_token_accuracy": 0.7524076342582703,
      "num_tokens": 1475469561.0,
      "step": 20530
    },
    {
      "entropy": 0.9029385939240455,
      "epoch": 5.598255655491959,
      "grad_norm": 0.10689301788806915,
      "learning_rate": 7.405929722454026e-05,
      "loss": 1.334,
      "mean_token_accuracy": 0.7467936128377914,
      "num_tokens": 1476172339.0,
      "step": 20540
    },
    {
      "entropy": 0.9055037021636962,
      "epoch": 5.600981193785772,
      "grad_norm": 0.1079697534441948,
      "learning_rate": 7.403295565640977e-05,
      "loss": 1.3405,
      "mean_token_accuracy": 0.7461986854672432,
      "num_tokens": 1476876827.0,
      "step": 20550
    },
    {
      "entropy": 0.9079792276024818,
      "epoch": 5.603706732079585,
      "grad_norm": 0.11904729902744293,
      "learning_rate": 7.40066054109532e-05,
      "loss": 1.3449,
      "mean_token_accuracy": 0.7490259379148483,
      "num_tokens": 1477585879.0,
      "step": 20560
    },
    {
      "entropy": 0.8998463958501816,
      "epoch": 5.606432270373399,
      "grad_norm": 0.1142992153763771,
      "learning_rate": 7.398024649768453e-05,
      "loss": 1.3209,
      "mean_token_accuracy": 0.7542578890919686,
      "num_tokens": 1478313059.0,
      "step": 20570
    },
    {
      "entropy": 0.9039091795682908,
      "epoch": 5.609157808667212,
      "grad_norm": 0.12061621248722076,
      "learning_rate": 7.395387892612092e-05,
      "loss": 1.3372,
      "mean_token_accuracy": 0.7542914986610413,
      "num_tokens": 1479043854.0,
      "step": 20580
    },
    {
      "entropy": 0.9103962168097496,
      "epoch": 5.611883346961025,
      "grad_norm": 0.11515385657548904,
      "learning_rate": 7.392750270578261e-05,
      "loss": 1.3387,
      "mean_token_accuracy": 0.7487710952758789,
      "num_tokens": 1479763616.0,
      "step": 20590
    },
    {
      "entropy": 0.9080936193466187,
      "epoch": 5.614608885254838,
      "grad_norm": 0.11225724965333939,
      "learning_rate": 7.3901117846193e-05,
      "loss": 1.3365,
      "mean_token_accuracy": 0.7493144363164902,
      "num_tokens": 1480487054.0,
      "step": 20600
    },
    {
      "entropy": 0.9011351138353347,
      "epoch": 5.617334423548651,
      "grad_norm": 0.11288349330425262,
      "learning_rate": 7.387472435687857e-05,
      "loss": 1.3302,
      "mean_token_accuracy": 0.7527704104781151,
      "num_tokens": 1481202839.0,
      "step": 20610
    },
    {
      "entropy": 0.9073983520269394,
      "epoch": 5.620059961842464,
      "grad_norm": 0.11644807457923889,
      "learning_rate": 7.384832224736897e-05,
      "loss": 1.34,
      "mean_token_accuracy": 0.7466869860887527,
      "num_tokens": 1481910828.0,
      "step": 20620
    },
    {
      "entropy": 0.9031661286950111,
      "epoch": 5.622785500136277,
      "grad_norm": 0.11115820705890656,
      "learning_rate": 7.382191152719691e-05,
      "loss": 1.3243,
      "mean_token_accuracy": 0.7502575770020485,
      "num_tokens": 1482639275.0,
      "step": 20630
    },
    {
      "entropy": 0.902911764383316,
      "epoch": 5.62551103843009,
      "grad_norm": 0.11243662238121033,
      "learning_rate": 7.379549220589823e-05,
      "loss": 1.3289,
      "mean_token_accuracy": 0.7518644094467163,
      "num_tokens": 1483362131.0,
      "step": 20640
    },
    {
      "entropy": 0.901501926779747,
      "epoch": 5.628236576723903,
      "grad_norm": 0.11614526808261871,
      "learning_rate": 7.376906429301189e-05,
      "loss": 1.3274,
      "mean_token_accuracy": 0.7536589324474334,
      "num_tokens": 1484092910.0,
      "step": 20650
    },
    {
      "entropy": 0.904544286429882,
      "epoch": 5.6309621150177165,
      "grad_norm": 0.11139136552810669,
      "learning_rate": 7.374262779807992e-05,
      "loss": 1.3318,
      "mean_token_accuracy": 0.7508685305714607,
      "num_tokens": 1484817903.0,
      "step": 20660
    },
    {
      "entropy": 0.9108938813209534,
      "epoch": 5.633687653311529,
      "grad_norm": 0.1144838035106659,
      "learning_rate": 7.371618273064749e-05,
      "loss": 1.339,
      "mean_token_accuracy": 0.7504678934812545,
      "num_tokens": 1485538554.0,
      "step": 20670
    },
    {
      "entropy": 0.9001557230949402,
      "epoch": 5.636413191605342,
      "grad_norm": 0.1142508015036583,
      "learning_rate": 7.368972910026282e-05,
      "loss": 1.3274,
      "mean_token_accuracy": 0.7570841550827027,
      "num_tokens": 1486267378.0,
      "step": 20680
    },
    {
      "entropy": 0.9114839643239975,
      "epoch": 5.639138729899155,
      "grad_norm": 0.1170973926782608,
      "learning_rate": 7.366326691647726e-05,
      "loss": 1.3408,
      "mean_token_accuracy": 0.7509129852056503,
      "num_tokens": 1486987586.0,
      "step": 20690
    },
    {
      "entropy": 0.9058082833886146,
      "epoch": 5.641864268192968,
      "grad_norm": 0.12553860247135162,
      "learning_rate": 7.363679618884521e-05,
      "loss": 1.331,
      "mean_token_accuracy": 0.7483681857585907,
      "num_tokens": 1487694735.0,
      "step": 20700
    },
    {
      "entropy": 0.9067086935043335,
      "epoch": 5.644589806486781,
      "grad_norm": 0.11567724496126175,
      "learning_rate": 7.36103169269242e-05,
      "loss": 1.3356,
      "mean_token_accuracy": 0.7496160626411438,
      "num_tokens": 1488407260.0,
      "step": 20710
    },
    {
      "entropy": 0.9018784865736962,
      "epoch": 5.647315344780594,
      "grad_norm": 0.11900515109300613,
      "learning_rate": 7.358382914027482e-05,
      "loss": 1.3223,
      "mean_token_accuracy": 0.7485745057463646,
      "num_tokens": 1489112368.0,
      "step": 20720
    },
    {
      "entropy": 0.9091901272535324,
      "epoch": 5.650040883074407,
      "grad_norm": 0.11961226910352707,
      "learning_rate": 7.355733283846072e-05,
      "loss": 1.336,
      "mean_token_accuracy": 0.7524806439876557,
      "num_tokens": 1489843077.0,
      "step": 20730
    },
    {
      "entropy": 0.9084666013717652,
      "epoch": 5.65276642136822,
      "grad_norm": 0.12814275920391083,
      "learning_rate": 7.353082803104864e-05,
      "loss": 1.3419,
      "mean_token_accuracy": 0.7467518895864487,
      "num_tokens": 1490539035.0,
      "step": 20740
    },
    {
      "entropy": 0.9046581968665123,
      "epoch": 5.655491959662033,
      "grad_norm": 0.11628738790750504,
      "learning_rate": 7.350431472760839e-05,
      "loss": 1.3355,
      "mean_token_accuracy": 0.7516518905758858,
      "num_tokens": 1491265803.0,
      "step": 20750
    },
    {
      "entropy": 0.9021990105509758,
      "epoch": 5.658217497955846,
      "grad_norm": 0.11885421723127365,
      "learning_rate": 7.347779293771287e-05,
      "loss": 1.3339,
      "mean_token_accuracy": 0.7534673601388931,
      "num_tokens": 1491991955.0,
      "step": 20760
    },
    {
      "entropy": 0.9049893140792846,
      "epoch": 5.660943036249659,
      "grad_norm": 0.1268477737903595,
      "learning_rate": 7.3451262670938e-05,
      "loss": 1.3386,
      "mean_token_accuracy": 0.7474268466234207,
      "num_tokens": 1492696757.0,
      "step": 20770
    },
    {
      "entropy": 0.9139819592237473,
      "epoch": 5.663668574543473,
      "grad_norm": 0.11796873807907104,
      "learning_rate": 7.34247239368628e-05,
      "loss": 1.3433,
      "mean_token_accuracy": 0.7465931549668312,
      "num_tokens": 1493421090.0,
      "step": 20780
    },
    {
      "entropy": 0.901647137105465,
      "epoch": 5.666394112837286,
      "grad_norm": 0.11307866871356964,
      "learning_rate": 7.339817674506934e-05,
      "loss": 1.3316,
      "mean_token_accuracy": 0.7515180230140686,
      "num_tokens": 1494140516.0,
      "step": 20790
    },
    {
      "entropy": 0.9086991220712661,
      "epoch": 5.669119651131099,
      "grad_norm": 0.10930800437927246,
      "learning_rate": 7.33716211051427e-05,
      "loss": 1.331,
      "mean_token_accuracy": 0.7510665923357009,
      "num_tokens": 1494856525.0,
      "step": 20800
    },
    {
      "entropy": 0.9052547365427017,
      "epoch": 5.671845189424912,
      "grad_norm": 0.1316278725862503,
      "learning_rate": 7.334505702667106e-05,
      "loss": 1.333,
      "mean_token_accuracy": 0.7491701155900955,
      "num_tokens": 1495568587.0,
      "step": 20810
    },
    {
      "entropy": 0.9048445269465446,
      "epoch": 5.6745707277187245,
      "grad_norm": 0.11319801211357117,
      "learning_rate": 7.331848451924565e-05,
      "loss": 1.3292,
      "mean_token_accuracy": 0.7502135202288628,
      "num_tokens": 1496282288.0,
      "step": 20820
    },
    {
      "entropy": 0.907095630466938,
      "epoch": 5.6772962660125375,
      "grad_norm": 0.11182509362697601,
      "learning_rate": 7.329190359246072e-05,
      "loss": 1.3322,
      "mean_token_accuracy": 0.7530362710356713,
      "num_tokens": 1497012885.0,
      "step": 20830
    },
    {
      "entropy": 0.9074946895241738,
      "epoch": 5.68002180430635,
      "grad_norm": 0.12122611701488495,
      "learning_rate": 7.326531425591353e-05,
      "loss": 1.3422,
      "mean_token_accuracy": 0.7459175392985344,
      "num_tokens": 1497710927.0,
      "step": 20840
    },
    {
      "entropy": 0.9031350791454316,
      "epoch": 5.682747342600163,
      "grad_norm": 0.10637669265270233,
      "learning_rate": 7.323871651920445e-05,
      "loss": 1.3348,
      "mean_token_accuracy": 0.7532668948173523,
      "num_tokens": 1498436997.0,
      "step": 20850
    },
    {
      "entropy": 0.8981655642390252,
      "epoch": 5.685472880893976,
      "grad_norm": 0.11300493776798248,
      "learning_rate": 7.321211039193683e-05,
      "loss": 1.328,
      "mean_token_accuracy": 0.750158154964447,
      "num_tokens": 1499138993.0,
      "step": 20860
    },
    {
      "entropy": 0.9084590286016464,
      "epoch": 5.688198419187789,
      "grad_norm": 0.1168494001030922,
      "learning_rate": 7.318549588371708e-05,
      "loss": 1.3322,
      "mean_token_accuracy": 0.7514884144067764,
      "num_tokens": 1499865175.0,
      "step": 20870
    },
    {
      "entropy": 0.9064156174659729,
      "epoch": 5.690923957481603,
      "grad_norm": 0.12878461182117462,
      "learning_rate": 7.315887300415461e-05,
      "loss": 1.3328,
      "mean_token_accuracy": 0.7508704990148545,
      "num_tokens": 1500585521.0,
      "step": 20880
    },
    {
      "entropy": 0.9077233120799064,
      "epoch": 5.693649495775416,
      "grad_norm": 0.11928346753120422,
      "learning_rate": 7.313224176286185e-05,
      "loss": 1.3326,
      "mean_token_accuracy": 0.7496184825897216,
      "num_tokens": 1501305914.0,
      "step": 20890
    },
    {
      "entropy": 0.9070781856775284,
      "epoch": 5.696375034069229,
      "grad_norm": 0.1107550859451294,
      "learning_rate": 7.310560216945427e-05,
      "loss": 1.3349,
      "mean_token_accuracy": 0.7506834283471108,
      "num_tokens": 1502023189.0,
      "step": 20900
    },
    {
      "entropy": 0.9059104502201081,
      "epoch": 5.699100572363042,
      "grad_norm": 0.11642283946275711,
      "learning_rate": 7.307895423355035e-05,
      "loss": 1.3336,
      "mean_token_accuracy": 0.7461674377322197,
      "num_tokens": 1502729027.0,
      "step": 20910
    },
    {
      "entropy": 0.9075036674737931,
      "epoch": 5.701826110656855,
      "grad_norm": 0.12099692225456238,
      "learning_rate": 7.305229796477158e-05,
      "loss": 1.3319,
      "mean_token_accuracy": 0.749399158358574,
      "num_tokens": 1503449220.0,
      "step": 20920
    },
    {
      "entropy": 0.9101551100611687,
      "epoch": 5.704551648950668,
      "grad_norm": 0.10946380347013474,
      "learning_rate": 7.302563337274246e-05,
      "loss": 1.3369,
      "mean_token_accuracy": 0.7491980984807014,
      "num_tokens": 1504151882.0,
      "step": 20930
    },
    {
      "entropy": 0.9068921029567718,
      "epoch": 5.707277187244481,
      "grad_norm": 0.1121559888124466,
      "learning_rate": 7.29989604670905e-05,
      "loss": 1.3357,
      "mean_token_accuracy": 0.7482832372188568,
      "num_tokens": 1504869987.0,
      "step": 20940
    },
    {
      "entropy": 0.907612769305706,
      "epoch": 5.710002725538294,
      "grad_norm": 0.10779725760221481,
      "learning_rate": 7.297227925744618e-05,
      "loss": 1.3368,
      "mean_token_accuracy": 0.7545360907912254,
      "num_tokens": 1505601375.0,
      "step": 20950
    },
    {
      "entropy": 0.9046850204467773,
      "epoch": 5.712728263832107,
      "grad_norm": 0.11786270141601562,
      "learning_rate": 7.294558975344301e-05,
      "loss": 1.3397,
      "mean_token_accuracy": 0.7510627821087837,
      "num_tokens": 1506323334.0,
      "step": 20960
    },
    {
      "entropy": 0.9153681427240372,
      "epoch": 5.71545380212592,
      "grad_norm": 0.11150839179754257,
      "learning_rate": 7.291889196471752e-05,
      "loss": 1.3361,
      "mean_token_accuracy": 0.7479392275214195,
      "num_tokens": 1507038388.0,
      "step": 20970
    },
    {
      "entropy": 0.913441763818264,
      "epoch": 5.7181793404197325,
      "grad_norm": 0.10878941416740417,
      "learning_rate": 7.289218590090916e-05,
      "loss": 1.3455,
      "mean_token_accuracy": 0.7470132410526276,
      "num_tokens": 1507751982.0,
      "step": 20980
    },
    {
      "entropy": 0.9085206687450409,
      "epoch": 5.7209048787135455,
      "grad_norm": 0.1076442077755928,
      "learning_rate": 7.286547157166041e-05,
      "loss": 1.3339,
      "mean_token_accuracy": 0.7513268738985062,
      "num_tokens": 1508484253.0,
      "step": 20990
    },
    {
      "entropy": 0.9020295903086663,
      "epoch": 5.723630417007359,
      "grad_norm": 0.11886076629161835,
      "learning_rate": 7.283874898661678e-05,
      "loss": 1.3304,
      "mean_token_accuracy": 0.7555002585053444,
      "num_tokens": 1509213233.0,
      "step": 21000
    },
    {
      "entropy": 0.9031096026301384,
      "epoch": 5.726355955301172,
      "grad_norm": 0.11685291677713394,
      "learning_rate": 7.281201815542666e-05,
      "loss": 1.3287,
      "mean_token_accuracy": 0.7478416725993157,
      "num_tokens": 1509922375.0,
      "step": 21010
    },
    {
      "entropy": 0.9087613329291344,
      "epoch": 5.729081493594985,
      "grad_norm": 0.11429954320192337,
      "learning_rate": 7.278527908774147e-05,
      "loss": 1.3377,
      "mean_token_accuracy": 0.7501318350434303,
      "num_tokens": 1510641545.0,
      "step": 21020
    },
    {
      "entropy": 0.9057684108614922,
      "epoch": 5.731807031888798,
      "grad_norm": 0.11861546337604523,
      "learning_rate": 7.275853179321565e-05,
      "loss": 1.332,
      "mean_token_accuracy": 0.7496027112007141,
      "num_tokens": 1511347859.0,
      "step": 21030
    },
    {
      "entropy": 0.9060276508331299,
      "epoch": 5.734532570182611,
      "grad_norm": 0.1182515025138855,
      "learning_rate": 7.273177628150652e-05,
      "loss": 1.3317,
      "mean_token_accuracy": 0.7510891646146775,
      "num_tokens": 1512070803.0,
      "step": 21040
    },
    {
      "entropy": 0.9113850757479668,
      "epoch": 5.737258108476424,
      "grad_norm": 0.11283525824546814,
      "learning_rate": 7.27050125622744e-05,
      "loss": 1.3345,
      "mean_token_accuracy": 0.7453511193394661,
      "num_tokens": 1512777940.0,
      "step": 21050
    },
    {
      "entropy": 0.9128487035632133,
      "epoch": 5.739983646770237,
      "grad_norm": 0.11141043901443481,
      "learning_rate": 7.267824064518264e-05,
      "loss": 1.3397,
      "mean_token_accuracy": 0.7464693874120713,
      "num_tokens": 1513490793.0,
      "step": 21060
    },
    {
      "entropy": 0.9017889350652695,
      "epoch": 5.74270918506405,
      "grad_norm": 0.1313389241695404,
      "learning_rate": 7.265146053989744e-05,
      "loss": 1.3375,
      "mean_token_accuracy": 0.7485280215740204,
      "num_tokens": 1514193471.0,
      "step": 21070
    },
    {
      "entropy": 0.9016789928078651,
      "epoch": 5.745434723357863,
      "grad_norm": 0.22550202906131744,
      "learning_rate": 7.262467225608805e-05,
      "loss": 1.331,
      "mean_token_accuracy": 0.7484898760914802,
      "num_tokens": 1514914647.0,
      "step": 21080
    },
    {
      "entropy": 0.9075780168175698,
      "epoch": 5.748160261651676,
      "grad_norm": 0.11743731051683426,
      "learning_rate": 7.259787580342662e-05,
      "loss": 1.3378,
      "mean_token_accuracy": 0.7536318570375442,
      "num_tokens": 1515649835.0,
      "step": 21090
    },
    {
      "entropy": 0.9135808631777763,
      "epoch": 5.75088579994549,
      "grad_norm": 0.1413869708776474,
      "learning_rate": 7.257107119158824e-05,
      "loss": 1.3457,
      "mean_token_accuracy": 0.7477648884057999,
      "num_tokens": 1516364830.0,
      "step": 21100
    },
    {
      "entropy": 0.9115987747907639,
      "epoch": 5.753611338239303,
      "grad_norm": 0.11431247740983963,
      "learning_rate": 7.2544258430251e-05,
      "loss": 1.3327,
      "mean_token_accuracy": 0.7489842206239701,
      "num_tokens": 1517075455.0,
      "step": 21110
    },
    {
      "entropy": 0.9034978196024894,
      "epoch": 5.7563368765331155,
      "grad_norm": 0.11637931317090988,
      "learning_rate": 7.251743752909589e-05,
      "loss": 1.3299,
      "mean_token_accuracy": 0.7487104117870331,
      "num_tokens": 1517787951.0,
      "step": 21120
    },
    {
      "entropy": 0.9044322043657302,
      "epoch": 5.7590624148269285,
      "grad_norm": 0.11377184092998505,
      "learning_rate": 7.249060849780684e-05,
      "loss": 1.3376,
      "mean_token_accuracy": 0.7492161333560944,
      "num_tokens": 1518487047.0,
      "step": 21130
    },
    {
      "entropy": 0.9142660677433014,
      "epoch": 5.761787953120741,
      "grad_norm": 0.1095757707953453,
      "learning_rate": 7.246377134607073e-05,
      "loss": 1.3423,
      "mean_token_accuracy": 0.7529979705810547,
      "num_tokens": 1519221826.0,
      "step": 21140
    },
    {
      "entropy": 0.9087027072906494,
      "epoch": 5.764513491414554,
      "grad_norm": 0.11647851765155792,
      "learning_rate": 7.24369260835774e-05,
      "loss": 1.3322,
      "mean_token_accuracy": 0.7485354587435722,
      "num_tokens": 1519942841.0,
      "step": 21150
    },
    {
      "entropy": 0.9132734730839729,
      "epoch": 5.767239029708367,
      "grad_norm": 0.11407783627510071,
      "learning_rate": 7.241007272001953e-05,
      "loss": 1.3449,
      "mean_token_accuracy": 0.7486987560987473,
      "num_tokens": 1520663817.0,
      "step": 21160
    },
    {
      "entropy": 0.9116878688335419,
      "epoch": 5.76996456800218,
      "grad_norm": 0.10810646414756775,
      "learning_rate": 7.238321126509282e-05,
      "loss": 1.3308,
      "mean_token_accuracy": 0.746232533454895,
      "num_tokens": 1521373149.0,
      "step": 21170
    },
    {
      "entropy": 0.9063387632369995,
      "epoch": 5.772690106295993,
      "grad_norm": 0.10843205451965332,
      "learning_rate": 7.235634172849584e-05,
      "loss": 1.3396,
      "mean_token_accuracy": 0.7472685724496841,
      "num_tokens": 1522086820.0,
      "step": 21180
    },
    {
      "entropy": 0.913464541733265,
      "epoch": 5.775415644589806,
      "grad_norm": 0.12286760658025742,
      "learning_rate": 7.232946411993009e-05,
      "loss": 1.3406,
      "mean_token_accuracy": 0.7466552630066872,
      "num_tokens": 1522801372.0,
      "step": 21190
    },
    {
      "entropy": 0.9062201827764511,
      "epoch": 5.778141182883619,
      "grad_norm": 0.1153421700000763,
      "learning_rate": 7.23025784491e-05,
      "loss": 1.3309,
      "mean_token_accuracy": 0.7528724387288094,
      "num_tokens": 1523528832.0,
      "step": 21200
    },
    {
      "entropy": 0.8964129894971847,
      "epoch": 5.780866721177432,
      "grad_norm": 0.11012628674507141,
      "learning_rate": 7.227568472571286e-05,
      "loss": 1.325,
      "mean_token_accuracy": 0.7586266621947289,
      "num_tokens": 1524272812.0,
      "step": 21210
    },
    {
      "entropy": 0.9096088528633117,
      "epoch": 5.783592259471246,
      "grad_norm": 0.14736290276050568,
      "learning_rate": 7.224878295947894e-05,
      "loss": 1.3503,
      "mean_token_accuracy": 0.7448491111397744,
      "num_tokens": 1524969510.0,
      "step": 21220
    },
    {
      "entropy": 0.8980430319905282,
      "epoch": 5.786317797765059,
      "grad_norm": 0.11593416333198547,
      "learning_rate": 7.222187316011138e-05,
      "loss": 1.331,
      "mean_token_accuracy": 0.7527055189013481,
      "num_tokens": 1525691152.0,
      "step": 21230
    },
    {
      "entropy": 0.8960934340953827,
      "epoch": 5.789043336058872,
      "grad_norm": 0.11741005629301071,
      "learning_rate": 7.219495533732618e-05,
      "loss": 1.3198,
      "mean_token_accuracy": 0.7602716162800789,
      "num_tokens": 1526428035.0,
      "step": 21240
    },
    {
      "entropy": 0.9068446651101112,
      "epoch": 5.791768874352685,
      "grad_norm": 0.10496514290571213,
      "learning_rate": 7.21680295008423e-05,
      "loss": 1.3359,
      "mean_token_accuracy": 0.7567587852478027,
      "num_tokens": 1527164602.0,
      "step": 21250
    },
    {
      "entropy": 0.9012785717844963,
      "epoch": 5.794494412646498,
      "grad_norm": 0.11333508789539337,
      "learning_rate": 7.214109566038158e-05,
      "loss": 1.3377,
      "mean_token_accuracy": 0.7537783280014991,
      "num_tokens": 1527883641.0,
      "step": 21260
    },
    {
      "entropy": 0.9095819979906082,
      "epoch": 5.797219950940311,
      "grad_norm": 0.11693883687257767,
      "learning_rate": 7.211415382566871e-05,
      "loss": 1.3361,
      "mean_token_accuracy": 0.7505184873938561,
      "num_tokens": 1528598023.0,
      "step": 21270
    },
    {
      "entropy": 0.9130471631884575,
      "epoch": 5.799945489234124,
      "grad_norm": 0.13102532923221588,
      "learning_rate": 7.208720400643131e-05,
      "loss": 1.3443,
      "mean_token_accuracy": 0.7470770180225372,
      "num_tokens": 1529307248.0,
      "step": 21280
    },
    {
      "entropy": 0.9062995433807373,
      "epoch": 5.8026710275279365,
      "grad_norm": 0.11358413845300674,
      "learning_rate": 7.206024621239986e-05,
      "loss": 1.3267,
      "mean_token_accuracy": 0.7522414743900299,
      "num_tokens": 1530032785.0,
      "step": 21290
    },
    {
      "entropy": 0.9074021801352501,
      "epoch": 5.8053965658217495,
      "grad_norm": 0.11888142675161362,
      "learning_rate": 7.203328045330773e-05,
      "loss": 1.3479,
      "mean_token_accuracy": 0.7494888335466385,
      "num_tokens": 1530747194.0,
      "step": 21300
    },
    {
      "entropy": 0.901043277978897,
      "epoch": 5.808122104115563,
      "grad_norm": 0.11507439613342285,
      "learning_rate": 7.200630673889117e-05,
      "loss": 1.334,
      "mean_token_accuracy": 0.7542605042457581,
      "num_tokens": 1531472972.0,
      "step": 21310
    },
    {
      "entropy": 0.9156588390469551,
      "epoch": 5.810847642409376,
      "grad_norm": 0.1293054074048996,
      "learning_rate": 7.197932507888929e-05,
      "loss": 1.3425,
      "mean_token_accuracy": 0.7489974901080132,
      "num_tokens": 1532187546.0,
      "step": 21320
    },
    {
      "entropy": 0.9082718387246131,
      "epoch": 5.813573180703189,
      "grad_norm": 0.11587195843458176,
      "learning_rate": 7.195233548304407e-05,
      "loss": 1.3319,
      "mean_token_accuracy": 0.7506336972117424,
      "num_tokens": 1532907614.0,
      "step": 21330
    },
    {
      "entropy": 0.90416921377182,
      "epoch": 5.816298718997002,
      "grad_norm": 0.11691489070653915,
      "learning_rate": 7.192533796110036e-05,
      "loss": 1.3473,
      "mean_token_accuracy": 0.7484967574477196,
      "num_tokens": 1533617716.0,
      "step": 21340
    },
    {
      "entropy": 0.907621382176876,
      "epoch": 5.819024257290815,
      "grad_norm": 0.11140576004981995,
      "learning_rate": 7.189833252280588e-05,
      "loss": 1.3337,
      "mean_token_accuracy": 0.7470261678099632,
      "num_tokens": 1534328518.0,
      "step": 21350
    },
    {
      "entropy": 0.9058356672525406,
      "epoch": 5.821749795584628,
      "grad_norm": 0.11089876294136047,
      "learning_rate": 7.187131917791118e-05,
      "loss": 1.3324,
      "mean_token_accuracy": 0.7501423612236977,
      "num_tokens": 1535042030.0,
      "step": 21360
    },
    {
      "entropy": 0.9052778214216233,
      "epoch": 5.824475333878441,
      "grad_norm": 0.11452427506446838,
      "learning_rate": 7.184429793616968e-05,
      "loss": 1.3285,
      "mean_token_accuracy": 0.7518876895308495,
      "num_tokens": 1535773704.0,
      "step": 21370
    },
    {
      "entropy": 0.9050400242209434,
      "epoch": 5.827200872172254,
      "grad_norm": 0.11301568895578384,
      "learning_rate": 7.18172688073377e-05,
      "loss": 1.3234,
      "mean_token_accuracy": 0.7526569321751595,
      "num_tokens": 1536499006.0,
      "step": 21380
    },
    {
      "entropy": 0.9012054517865181,
      "epoch": 5.829926410466067,
      "grad_norm": 0.11559009552001953,
      "learning_rate": 7.179023180117433e-05,
      "loss": 1.332,
      "mean_token_accuracy": 0.7539320036768913,
      "num_tokens": 1537223155.0,
      "step": 21390
    },
    {
      "entropy": 0.896777668595314,
      "epoch": 5.83265194875988,
      "grad_norm": 0.11815473437309265,
      "learning_rate": 7.176318692744153e-05,
      "loss": 1.3294,
      "mean_token_accuracy": 0.7591805055737495,
      "num_tokens": 1537956904.0,
      "step": 21400
    },
    {
      "entropy": 0.8999134451150894,
      "epoch": 5.835377487053693,
      "grad_norm": 0.1185450404882431,
      "learning_rate": 7.173613419590411e-05,
      "loss": 1.3261,
      "mean_token_accuracy": 0.7577306047081948,
      "num_tokens": 1538689494.0,
      "step": 21410
    },
    {
      "entropy": 0.899851243197918,
      "epoch": 5.838103025347506,
      "grad_norm": 0.11542113125324249,
      "learning_rate": 7.170907361632972e-05,
      "loss": 1.3286,
      "mean_token_accuracy": 0.7565847590565682,
      "num_tokens": 1539412090.0,
      "step": 21420
    },
    {
      "entropy": 0.9017447680234909,
      "epoch": 5.8408285636413195,
      "grad_norm": 0.11492068320512772,
      "learning_rate": 7.168200519848884e-05,
      "loss": 1.3366,
      "mean_token_accuracy": 0.7529019027948379,
      "num_tokens": 1540132195.0,
      "step": 21430
    },
    {
      "entropy": 0.9049430832266807,
      "epoch": 5.8435541019351325,
      "grad_norm": 0.11647587269544601,
      "learning_rate": 7.165492895215481e-05,
      "loss": 1.3325,
      "mean_token_accuracy": 0.7481189534068108,
      "num_tokens": 1540835722.0,
      "step": 21440
    },
    {
      "entropy": 0.8994492426514625,
      "epoch": 5.846279640228945,
      "grad_norm": 0.11033880710601807,
      "learning_rate": 7.16278448871037e-05,
      "loss": 1.3299,
      "mean_token_accuracy": 0.7494186624884606,
      "num_tokens": 1541553860.0,
      "step": 21450
    },
    {
      "entropy": 0.9022593319416046,
      "epoch": 5.849005178522758,
      "grad_norm": 0.10914087295532227,
      "learning_rate": 7.160075301311451e-05,
      "loss": 1.3307,
      "mean_token_accuracy": 0.7485667675733566,
      "num_tokens": 1542260817.0,
      "step": 21460
    },
    {
      "entropy": 0.8925522163510322,
      "epoch": 5.851730716816571,
      "grad_norm": 0.10634701699018478,
      "learning_rate": 7.157365333996903e-05,
      "loss": 1.3217,
      "mean_token_accuracy": 0.7552912056446075,
      "num_tokens": 1542986865.0,
      "step": 21470
    },
    {
      "entropy": 0.896972981095314,
      "epoch": 5.854456255110384,
      "grad_norm": 0.11877600848674774,
      "learning_rate": 7.154654587745181e-05,
      "loss": 1.326,
      "mean_token_accuracy": 0.7548259600996972,
      "num_tokens": 1543711426.0,
      "step": 21480
    },
    {
      "entropy": 0.9042829141020775,
      "epoch": 5.857181793404197,
      "grad_norm": 0.10897119343280792,
      "learning_rate": 7.151943063535029e-05,
      "loss": 1.332,
      "mean_token_accuracy": 0.7512248262763024,
      "num_tokens": 1544447293.0,
      "step": 21490
    },
    {
      "entropy": 0.9050748169422149,
      "epoch": 5.85990733169801,
      "grad_norm": 0.11239845305681229,
      "learning_rate": 7.149230762345466e-05,
      "loss": 1.3252,
      "mean_token_accuracy": 0.7527645841240883,
      "num_tokens": 1545177492.0,
      "step": 21500
    },
    {
      "entropy": 0.9117373630404473,
      "epoch": 5.862632869991823,
      "grad_norm": 0.11432221531867981,
      "learning_rate": 7.146517685155797e-05,
      "loss": 1.3329,
      "mean_token_accuracy": 0.7515347197651863,
      "num_tokens": 1545910253.0,
      "step": 21510
    },
    {
      "entropy": 0.9061395451426506,
      "epoch": 5.865358408285636,
      "grad_norm": 0.12396573275327682,
      "learning_rate": 7.143803832945601e-05,
      "loss": 1.3271,
      "mean_token_accuracy": 0.7548024579882622,
      "num_tokens": 1546638251.0,
      "step": 21520
    },
    {
      "entropy": 0.9042191088199616,
      "epoch": 5.86808394657945,
      "grad_norm": 0.1185867041349411,
      "learning_rate": 7.141089206694743e-05,
      "loss": 1.3337,
      "mean_token_accuracy": 0.7500158175826073,
      "num_tokens": 1547351455.0,
      "step": 21530
    },
    {
      "entropy": 0.9097189202904701,
      "epoch": 5.870809484873263,
      "grad_norm": 0.11794767528772354,
      "learning_rate": 7.138373807383362e-05,
      "loss": 1.3356,
      "mean_token_accuracy": 0.7525267764925957,
      "num_tokens": 1548075104.0,
      "step": 21540
    },
    {
      "entropy": 0.9049504280090332,
      "epoch": 5.873535023167076,
      "grad_norm": 0.12269119918346405,
      "learning_rate": 7.13565763599188e-05,
      "loss": 1.3338,
      "mean_token_accuracy": 0.7550853103399277,
      "num_tokens": 1548807941.0,
      "step": 21550
    },
    {
      "entropy": 0.9065029069781303,
      "epoch": 5.876260561460889,
      "grad_norm": 0.1095624566078186,
      "learning_rate": 7.132940693500998e-05,
      "loss": 1.3383,
      "mean_token_accuracy": 0.7514843955636025,
      "num_tokens": 1549522912.0,
      "step": 21560
    },
    {
      "entropy": 0.9028425559401512,
      "epoch": 5.878986099754702,
      "grad_norm": 0.11521097272634506,
      "learning_rate": 7.13022298089169e-05,
      "loss": 1.3322,
      "mean_token_accuracy": 0.752574360370636,
      "num_tokens": 1550251480.0,
      "step": 21570
    },
    {
      "entropy": 0.9045530989766121,
      "epoch": 5.881711638048515,
      "grad_norm": 0.10882606357336044,
      "learning_rate": 7.127504499145213e-05,
      "loss": 1.3189,
      "mean_token_accuracy": 0.7573057830333709,
      "num_tokens": 1550997916.0,
      "step": 21580
    },
    {
      "entropy": 0.8961788237094879,
      "epoch": 5.8844371763423275,
      "grad_norm": 0.10879376530647278,
      "learning_rate": 7.124785249243103e-05,
      "loss": 1.3243,
      "mean_token_accuracy": 0.7560637041926384,
      "num_tokens": 1551737601.0,
      "step": 21590
    },
    {
      "entropy": 0.9038919240236283,
      "epoch": 5.8871627146361405,
      "grad_norm": 0.12050269544124603,
      "learning_rate": 7.12206523216717e-05,
      "loss": 1.3271,
      "mean_token_accuracy": 0.7502954587340355,
      "num_tokens": 1552448903.0,
      "step": 21600
    },
    {
      "entropy": 0.8999939456582069,
      "epoch": 5.889888252929953,
      "grad_norm": 0.12266591191291809,
      "learning_rate": 7.119344448899498e-05,
      "loss": 1.3241,
      "mean_token_accuracy": 0.7521205097436905,
      "num_tokens": 1553167664.0,
      "step": 21610
    },
    {
      "entropy": 0.9063843786716461,
      "epoch": 5.892613791223766,
      "grad_norm": 0.11235978454351425,
      "learning_rate": 7.116622900422456e-05,
      "loss": 1.3351,
      "mean_token_accuracy": 0.7485270857810974,
      "num_tokens": 1553874680.0,
      "step": 21620
    },
    {
      "entropy": 0.9047386437654495,
      "epoch": 5.895339329517579,
      "grad_norm": 0.12159693241119385,
      "learning_rate": 7.113900587718685e-05,
      "loss": 1.3374,
      "mean_token_accuracy": 0.7542475134134292,
      "num_tokens": 1554606300.0,
      "step": 21630
    },
    {
      "entropy": 0.9033267572522163,
      "epoch": 5.898064867811392,
      "grad_norm": 0.12074053287506104,
      "learning_rate": 7.111177511771099e-05,
      "loss": 1.3328,
      "mean_token_accuracy": 0.7475815027952194,
      "num_tokens": 1555307808.0,
      "step": 21640
    },
    {
      "entropy": 0.9042378276586532,
      "epoch": 5.900790406105206,
      "grad_norm": 0.11893583834171295,
      "learning_rate": 7.108453673562891e-05,
      "loss": 1.3309,
      "mean_token_accuracy": 0.7536738410592079,
      "num_tokens": 1556030679.0,
      "step": 21650
    },
    {
      "entropy": 0.9091543048620224,
      "epoch": 5.903515944399019,
      "grad_norm": 0.11554117500782013,
      "learning_rate": 7.10572907407753e-05,
      "loss": 1.3421,
      "mean_token_accuracy": 0.7465230345726013,
      "num_tokens": 1556730101.0,
      "step": 21660
    },
    {
      "entropy": 0.9006141796708107,
      "epoch": 5.906241482692832,
      "grad_norm": 0.1543373018503189,
      "learning_rate": 7.103003714298757e-05,
      "loss": 1.3273,
      "mean_token_accuracy": 0.752702134847641,
      "num_tokens": 1557444362.0,
      "step": 21670
    },
    {
      "entropy": 0.9043982520699501,
      "epoch": 5.908967020986645,
      "grad_norm": 0.12153176963329315,
      "learning_rate": 7.100277595210591e-05,
      "loss": 1.3331,
      "mean_token_accuracy": 0.7517162814736367,
      "num_tokens": 1558174426.0,
      "step": 21680
    },
    {
      "entropy": 0.903817056119442,
      "epoch": 5.911692559280458,
      "grad_norm": 0.11969004571437836,
      "learning_rate": 7.097550717797321e-05,
      "loss": 1.3314,
      "mean_token_accuracy": 0.7550696551799774,
      "num_tokens": 1558914880.0,
      "step": 21690
    },
    {
      "entropy": 0.9011016815900803,
      "epoch": 5.914418097574271,
      "grad_norm": 0.12317018210887909,
      "learning_rate": 7.094823083043511e-05,
      "loss": 1.324,
      "mean_token_accuracy": 0.7550008714199066,
      "num_tokens": 1559646795.0,
      "step": 21700
    },
    {
      "entropy": 0.91651850938797,
      "epoch": 5.917143635868084,
      "grad_norm": 0.11649711430072784,
      "learning_rate": 7.092094691934004e-05,
      "loss": 1.3401,
      "mean_token_accuracy": 0.7480064913630485,
      "num_tokens": 1560372629.0,
      "step": 21710
    },
    {
      "entropy": 0.9046705111861228,
      "epoch": 5.919869174161897,
      "grad_norm": 0.12039873003959656,
      "learning_rate": 7.089365545453906e-05,
      "loss": 1.3261,
      "mean_token_accuracy": 0.7540947735309601,
      "num_tokens": 1561103426.0,
      "step": 21720
    },
    {
      "entropy": 0.9046516075730324,
      "epoch": 5.92259471245571,
      "grad_norm": 0.11867029964923859,
      "learning_rate": 7.086635644588604e-05,
      "loss": 1.3447,
      "mean_token_accuracy": 0.7449310541152954,
      "num_tokens": 1561803039.0,
      "step": 21730
    },
    {
      "entropy": 0.9128703400492668,
      "epoch": 5.9253202507495235,
      "grad_norm": 0.10717029124498367,
      "learning_rate": 7.083904990323755e-05,
      "loss": 1.3421,
      "mean_token_accuracy": 0.7493982627987862,
      "num_tokens": 1562528882.0,
      "step": 21740
    },
    {
      "entropy": 0.9036580204963685,
      "epoch": 5.9280457890433365,
      "grad_norm": 0.12170601636171341,
      "learning_rate": 7.081173583645286e-05,
      "loss": 1.3242,
      "mean_token_accuracy": 0.7546490073204041,
      "num_tokens": 1563254856.0,
      "step": 21750
    },
    {
      "entropy": 0.9064361125230789,
      "epoch": 5.930771327337149,
      "grad_norm": 0.11220008134841919,
      "learning_rate": 7.078441425539398e-05,
      "loss": 1.3402,
      "mean_token_accuracy": 0.7528118431568146,
      "num_tokens": 1563972881.0,
      "step": 21760
    },
    {
      "entropy": 0.9053695395588874,
      "epoch": 5.933496865630962,
      "grad_norm": 0.11225832998752594,
      "learning_rate": 7.075708516992562e-05,
      "loss": 1.3287,
      "mean_token_accuracy": 0.7548480466008186,
      "num_tokens": 1564702125.0,
      "step": 21770
    },
    {
      "entropy": 0.904731398820877,
      "epoch": 5.936222403924775,
      "grad_norm": 0.12214484065771103,
      "learning_rate": 7.07297485899152e-05,
      "loss": 1.3336,
      "mean_token_accuracy": 0.7515589758753777,
      "num_tokens": 1565429318.0,
      "step": 21780
    },
    {
      "entropy": 0.9099110662937164,
      "epoch": 5.938947942218588,
      "grad_norm": 0.13402391970157623,
      "learning_rate": 7.070240452523288e-05,
      "loss": 1.3463,
      "mean_token_accuracy": 0.7511887595057487,
      "num_tokens": 1566148102.0,
      "step": 21790
    },
    {
      "entropy": 0.9190353766083718,
      "epoch": 5.941673480512401,
      "grad_norm": 0.1147463247179985,
      "learning_rate": 7.067505298575146e-05,
      "loss": 1.3533,
      "mean_token_accuracy": 0.7461867153644561,
      "num_tokens": 1566857817.0,
      "step": 21800
    },
    {
      "entropy": 0.909945385158062,
      "epoch": 5.944399018806214,
      "grad_norm": 0.11669431626796722,
      "learning_rate": 7.064769398134649e-05,
      "loss": 1.3383,
      "mean_token_accuracy": 0.7492309078574181,
      "num_tokens": 1567575204.0,
      "step": 21810
    },
    {
      "entropy": 0.9009095013141633,
      "epoch": 5.947124557100027,
      "grad_norm": 0.1547296643257141,
      "learning_rate": 7.062032752189619e-05,
      "loss": 1.3304,
      "mean_token_accuracy": 0.7499802619218826,
      "num_tokens": 1568279167.0,
      "step": 21820
    },
    {
      "entropy": 0.8945521906018257,
      "epoch": 5.94985009539384,
      "grad_norm": 0.11645699292421341,
      "learning_rate": 7.059295361728149e-05,
      "loss": 1.3376,
      "mean_token_accuracy": 0.7529954999685288,
      "num_tokens": 1568989407.0,
      "step": 21830
    },
    {
      "entropy": 0.9087991073727608,
      "epoch": 5.952575633687653,
      "grad_norm": 0.11843431740999222,
      "learning_rate": 7.056557227738597e-05,
      "loss": 1.3401,
      "mean_token_accuracy": 0.7492764741182327,
      "num_tokens": 1569703038.0,
      "step": 21840
    },
    {
      "entropy": 0.9076632171869278,
      "epoch": 5.955301171981466,
      "grad_norm": 0.1186591237783432,
      "learning_rate": 7.053818351209596e-05,
      "loss": 1.3385,
      "mean_token_accuracy": 0.7492095783352852,
      "num_tokens": 1570419352.0,
      "step": 21850
    },
    {
      "entropy": 0.9066355928778649,
      "epoch": 5.95802671027528,
      "grad_norm": 0.11073645204305649,
      "learning_rate": 7.05107873313004e-05,
      "loss": 1.332,
      "mean_token_accuracy": 0.7488937497138977,
      "num_tokens": 1571133758.0,
      "step": 21860
    },
    {
      "entropy": 0.9008475184440613,
      "epoch": 5.960752248569093,
      "grad_norm": 0.11767814308404922,
      "learning_rate": 7.048338374489096e-05,
      "loss": 1.3276,
      "mean_token_accuracy": 0.7530233487486839,
      "num_tokens": 1571865638.0,
      "step": 21870
    },
    {
      "entropy": 0.9052358761429786,
      "epoch": 5.963477786862906,
      "grad_norm": 0.15197525918483734,
      "learning_rate": 7.045597276276197e-05,
      "loss": 1.3222,
      "mean_token_accuracy": 0.7525520205497742,
      "num_tokens": 1572599090.0,
      "step": 21880
    },
    {
      "entropy": 0.8935629814863205,
      "epoch": 5.966203325156719,
      "grad_norm": 0.12846499681472778,
      "learning_rate": 7.04285543948104e-05,
      "loss": 1.3214,
      "mean_token_accuracy": 0.754536534845829,
      "num_tokens": 1573322068.0,
      "step": 21890
    },
    {
      "entropy": 0.8992321044206619,
      "epoch": 5.9689288634505315,
      "grad_norm": 0.12908099591732025,
      "learning_rate": 7.040112865093593e-05,
      "loss": 1.3343,
      "mean_token_accuracy": 0.7531968131661415,
      "num_tokens": 1574043501.0,
      "step": 21900
    },
    {
      "entropy": 0.9084354728460312,
      "epoch": 5.9716544017443445,
      "grad_norm": 0.11236303299665451,
      "learning_rate": 7.037369554104087e-05,
      "loss": 1.3346,
      "mean_token_accuracy": 0.7484549596905709,
      "num_tokens": 1574760565.0,
      "step": 21910
    },
    {
      "entropy": 0.913992977142334,
      "epoch": 5.974379940038157,
      "grad_norm": 0.11477797478437424,
      "learning_rate": 7.034625507503022e-05,
      "loss": 1.3418,
      "mean_token_accuracy": 0.7486337691545486,
      "num_tokens": 1575480198.0,
      "step": 21920
    },
    {
      "entropy": 0.9076977044343948,
      "epoch": 5.97710547833197,
      "grad_norm": 0.1153983324766159,
      "learning_rate": 7.031880726281161e-05,
      "loss": 1.3283,
      "mean_token_accuracy": 0.7501808911561966,
      "num_tokens": 1576202011.0,
      "step": 21930
    },
    {
      "entropy": 0.9015500947833062,
      "epoch": 5.979831016625783,
      "grad_norm": 0.10736705362796783,
      "learning_rate": 7.029135211429534e-05,
      "loss": 1.3306,
      "mean_token_accuracy": 0.7478416576981545,
      "num_tokens": 1576908139.0,
      "step": 21940
    },
    {
      "entropy": 0.9075870469212532,
      "epoch": 5.982556554919596,
      "grad_norm": 0.11770978569984436,
      "learning_rate": 7.026388963939435e-05,
      "loss": 1.3385,
      "mean_token_accuracy": 0.7497616693377495,
      "num_tokens": 1577616701.0,
      "step": 21950
    },
    {
      "entropy": 0.9064632877707481,
      "epoch": 5.98528209321341,
      "grad_norm": 0.12228012084960938,
      "learning_rate": 7.023641984802421e-05,
      "loss": 1.3295,
      "mean_token_accuracy": 0.7563860073685647,
      "num_tokens": 1578352596.0,
      "step": 21960
    },
    {
      "entropy": 0.9038796693086624,
      "epoch": 5.988007631507223,
      "grad_norm": 0.10734150558710098,
      "learning_rate": 7.020894275010318e-05,
      "loss": 1.3345,
      "mean_token_accuracy": 0.7479429483413697,
      "num_tokens": 1579065080.0,
      "step": 21970
    },
    {
      "entropy": 0.9043041512370109,
      "epoch": 5.990733169801036,
      "grad_norm": 0.11305051296949387,
      "learning_rate": 7.018145835555209e-05,
      "loss": 1.3278,
      "mean_token_accuracy": 0.7520740538835525,
      "num_tokens": 1579785660.0,
      "step": 21980
    },
    {
      "entropy": 0.9011677592992783,
      "epoch": 5.993458708094849,
      "grad_norm": 0.1225271001458168,
      "learning_rate": 7.015396667429448e-05,
      "loss": 1.3272,
      "mean_token_accuracy": 0.7554933309555054,
      "num_tokens": 1580508583.0,
      "step": 21990
    },
    {
      "entropy": 0.9020877957344056,
      "epoch": 5.996184246388662,
      "grad_norm": 0.11489120870828629,
      "learning_rate": 7.012646771625644e-05,
      "loss": 1.3282,
      "mean_token_accuracy": 0.7508908599615097,
      "num_tokens": 1581217757.0,
      "step": 22000
    },
    {
      "entropy": 0.9066957876086235,
      "epoch": 5.998909784682475,
      "grad_norm": 0.10596401244401932,
      "learning_rate": 7.009896149136674e-05,
      "loss": 1.3355,
      "mean_token_accuracy": 0.7449407547712326,
      "num_tokens": 1581915021.0,
      "step": 22010
    },
    {
      "entropy": 0.8996339097619057,
      "epoch": 6.001635322976288,
      "grad_norm": 0.11777181178331375,
      "learning_rate": 7.00714480095568e-05,
      "loss": 1.3146,
      "mean_token_accuracy": 0.7546662047505379,
      "num_tokens": 1582646097.0,
      "step": 22020
    },
    {
      "entropy": 0.8918565154075623,
      "epoch": 6.004360861270101,
      "grad_norm": 0.12640653550624847,
      "learning_rate": 7.004392728076058e-05,
      "loss": 1.3255,
      "mean_token_accuracy": 0.7498223036527634,
      "num_tokens": 1583355899.0,
      "step": 22030
    },
    {
      "entropy": 0.8883702427148819,
      "epoch": 6.007086399563914,
      "grad_norm": 0.13016051054000854,
      "learning_rate": 7.00163993149147e-05,
      "loss": 1.3195,
      "mean_token_accuracy": 0.7538187056779861,
      "num_tokens": 1584075326.0,
      "step": 22040
    },
    {
      "entropy": 0.8920476689934731,
      "epoch": 6.009811937857727,
      "grad_norm": 0.12302690744400024,
      "learning_rate": 6.998886412195842e-05,
      "loss": 1.3202,
      "mean_token_accuracy": 0.7489654153585434,
      "num_tokens": 1584771151.0,
      "step": 22050
    },
    {
      "entropy": 0.8842059820890427,
      "epoch": 6.0125374761515396,
      "grad_norm": 0.12850545346736908,
      "learning_rate": 6.996132171183358e-05,
      "loss": 1.3142,
      "mean_token_accuracy": 0.7590628907084465,
      "num_tokens": 1585492023.0,
      "step": 22060
    },
    {
      "entropy": 0.8875583812594414,
      "epoch": 6.015263014445353,
      "grad_norm": 0.13881002366542816,
      "learning_rate": 6.99337720944846e-05,
      "loss": 1.3087,
      "mean_token_accuracy": 0.7540833830833436,
      "num_tokens": 1586216570.0,
      "step": 22070
    },
    {
      "entropy": 0.8929170578718185,
      "epoch": 6.017988552739166,
      "grad_norm": 0.1315547227859497,
      "learning_rate": 6.990621527985856e-05,
      "loss": 1.3195,
      "mean_token_accuracy": 0.7589869886636734,
      "num_tokens": 1586953873.0,
      "step": 22080
    },
    {
      "entropy": 0.8922048270702362,
      "epoch": 6.020714091032979,
      "grad_norm": 0.11654641479253769,
      "learning_rate": 6.987865127790508e-05,
      "loss": 1.3128,
      "mean_token_accuracy": 0.7577459946274757,
      "num_tokens": 1587678683.0,
      "step": 22090
    },
    {
      "entropy": 0.8893833413720131,
      "epoch": 6.023439629326792,
      "grad_norm": 0.11815489083528519,
      "learning_rate": 6.985108009857643e-05,
      "loss": 1.3199,
      "mean_token_accuracy": 0.7532740905880928,
      "num_tokens": 1588396643.0,
      "step": 22100
    },
    {
      "entropy": 0.890563715994358,
      "epoch": 6.026165167620605,
      "grad_norm": 0.12686485052108765,
      "learning_rate": 6.982350175182744e-05,
      "loss": 1.3157,
      "mean_token_accuracy": 0.7522914201021195,
      "num_tokens": 1589108260.0,
      "step": 22110
    },
    {
      "entropy": 0.8936678737401962,
      "epoch": 6.028890705914418,
      "grad_norm": 0.1292998045682907,
      "learning_rate": 6.979591624761552e-05,
      "loss": 1.3214,
      "mean_token_accuracy": 0.7485388427972793,
      "num_tokens": 1589817771.0,
      "step": 22120
    },
    {
      "entropy": 0.8897582158446312,
      "epoch": 6.031616244208231,
      "grad_norm": 0.12712526321411133,
      "learning_rate": 6.976832359590067e-05,
      "loss": 1.321,
      "mean_token_accuracy": 0.7518042385578155,
      "num_tokens": 1590534431.0,
      "step": 22130
    },
    {
      "entropy": 0.8897338256239891,
      "epoch": 6.034341782502044,
      "grad_norm": 0.13390809297561646,
      "learning_rate": 6.974072380664551e-05,
      "loss": 1.3229,
      "mean_token_accuracy": 0.7534710794687272,
      "num_tokens": 1591246909.0,
      "step": 22140
    },
    {
      "entropy": 0.8840497434139252,
      "epoch": 6.037067320795857,
      "grad_norm": 0.12325893342494965,
      "learning_rate": 6.971311688981518e-05,
      "loss": 1.3133,
      "mean_token_accuracy": 0.7563563913106919,
      "num_tokens": 1591968830.0,
      "step": 22150
    },
    {
      "entropy": 0.8909603416919708,
      "epoch": 6.03979285908967,
      "grad_norm": 0.13583821058273315,
      "learning_rate": 6.968550285537745e-05,
      "loss": 1.3184,
      "mean_token_accuracy": 0.7549219325184822,
      "num_tokens": 1592680889.0,
      "step": 22160
    },
    {
      "entropy": 0.8920790418982506,
      "epoch": 6.042518397383483,
      "grad_norm": 0.1283435970544815,
      "learning_rate": 6.965788171330258e-05,
      "loss": 1.3215,
      "mean_token_accuracy": 0.7514332816004753,
      "num_tokens": 1593394633.0,
      "step": 22170
    },
    {
      "entropy": 0.8921322748064995,
      "epoch": 6.045243935677297,
      "grad_norm": 0.127129465341568,
      "learning_rate": 6.963025347356349e-05,
      "loss": 1.3206,
      "mean_token_accuracy": 0.7548703938722611,
      "num_tokens": 1594118725.0,
      "step": 22180
    },
    {
      "entropy": 0.8960848644375801,
      "epoch": 6.04796947397111,
      "grad_norm": 0.12366831302642822,
      "learning_rate": 6.960261814613558e-05,
      "loss": 1.3234,
      "mean_token_accuracy": 0.7518098592758179,
      "num_tokens": 1594840620.0,
      "step": 22190
    },
    {
      "entropy": 0.8972969114780426,
      "epoch": 6.050695012264923,
      "grad_norm": 0.12490490078926086,
      "learning_rate": 6.957497574099687e-05,
      "loss": 1.3245,
      "mean_token_accuracy": 0.7537968784570694,
      "num_tokens": 1595560984.0,
      "step": 22200
    },
    {
      "entropy": 0.8910147756338119,
      "epoch": 6.0534205505587355,
      "grad_norm": 0.12830320000648499,
      "learning_rate": 6.95473262681279e-05,
      "loss": 1.3071,
      "mean_token_accuracy": 0.753321748971939,
      "num_tokens": 1596278756.0,
      "step": 22210
    },
    {
      "entropy": 0.8912463679909706,
      "epoch": 6.0561460888525485,
      "grad_norm": 0.12675067782402039,
      "learning_rate": 6.951966973751178e-05,
      "loss": 1.3158,
      "mean_token_accuracy": 0.7535087749361992,
      "num_tokens": 1596995412.0,
      "step": 22220
    },
    {
      "entropy": 0.886512254178524,
      "epoch": 6.058871627146361,
      "grad_norm": 0.12018350511789322,
      "learning_rate": 6.949200615913417e-05,
      "loss": 1.3167,
      "mean_token_accuracy": 0.7514047920703888,
      "num_tokens": 1597692120.0,
      "step": 22230
    },
    {
      "entropy": 0.8892407730221749,
      "epoch": 6.061597165440174,
      "grad_norm": 0.1190294548869133,
      "learning_rate": 6.946433554298326e-05,
      "loss": 1.3115,
      "mean_token_accuracy": 0.7567781344056129,
      "num_tokens": 1598422283.0,
      "step": 22240
    },
    {
      "entropy": 0.8828908443450928,
      "epoch": 6.064322703733987,
      "grad_norm": 0.12898795306682587,
      "learning_rate": 6.943665789904979e-05,
      "loss": 1.3104,
      "mean_token_accuracy": 0.7551097676157952,
      "num_tokens": 1599143476.0,
      "step": 22250
    },
    {
      "entropy": 0.9043629035353661,
      "epoch": 6.0670482420278,
      "grad_norm": 0.12606586515903473,
      "learning_rate": 6.940897323732703e-05,
      "loss": 1.3178,
      "mean_token_accuracy": 0.7502626135945321,
      "num_tokens": 1599866042.0,
      "step": 22260
    },
    {
      "entropy": 0.8884061798453331,
      "epoch": 6.069773780321613,
      "grad_norm": 0.12459734082221985,
      "learning_rate": 6.938128156781078e-05,
      "loss": 1.3144,
      "mean_token_accuracy": 0.753172567486763,
      "num_tokens": 1600580503.0,
      "step": 22270
    },
    {
      "entropy": 0.8842171415686607,
      "epoch": 6.072499318615426,
      "grad_norm": 0.12710915505886078,
      "learning_rate": 6.935358290049939e-05,
      "loss": 1.3119,
      "mean_token_accuracy": 0.7543427735567093,
      "num_tokens": 1601299172.0,
      "step": 22280
    },
    {
      "entropy": 0.8847817584872246,
      "epoch": 6.07522485690924,
      "grad_norm": 0.12989383935928345,
      "learning_rate": 6.932587724539374e-05,
      "loss": 1.3106,
      "mean_token_accuracy": 0.7561130687594414,
      "num_tokens": 1602015327.0,
      "step": 22290
    },
    {
      "entropy": 0.8922547936439514,
      "epoch": 6.077950395203053,
      "grad_norm": 0.14769838750362396,
      "learning_rate": 6.929816461249719e-05,
      "loss": 1.3194,
      "mean_token_accuracy": 0.7522080436348915,
      "num_tokens": 1602722231.0,
      "step": 22300
    },
    {
      "entropy": 0.8895033150911331,
      "epoch": 6.080675933496866,
      "grad_norm": 0.13418523967266083,
      "learning_rate": 6.927044501181568e-05,
      "loss": 1.313,
      "mean_token_accuracy": 0.7518047258257866,
      "num_tokens": 1603431915.0,
      "step": 22310
    },
    {
      "entropy": 0.8931319579482079,
      "epoch": 6.083401471790679,
      "grad_norm": 0.12505514919757843,
      "learning_rate": 6.92427184533576e-05,
      "loss": 1.3205,
      "mean_token_accuracy": 0.7564160689711571,
      "num_tokens": 1604167329.0,
      "step": 22320
    },
    {
      "entropy": 0.8987172707915306,
      "epoch": 6.086127010084492,
      "grad_norm": 0.1262582689523697,
      "learning_rate": 6.921498494713392e-05,
      "loss": 1.3306,
      "mean_token_accuracy": 0.7501651674509049,
      "num_tokens": 1604880356.0,
      "step": 22330
    },
    {
      "entropy": 0.8913463935256004,
      "epoch": 6.088852548378305,
      "grad_norm": 0.13527299463748932,
      "learning_rate": 6.918724450315805e-05,
      "loss": 1.3199,
      "mean_token_accuracy": 0.7520601511001587,
      "num_tokens": 1605594598.0,
      "step": 22340
    },
    {
      "entropy": 0.8869917944073678,
      "epoch": 6.091578086672118,
      "grad_norm": 0.12402289360761642,
      "learning_rate": 6.915949713144599e-05,
      "loss": 1.3152,
      "mean_token_accuracy": 0.7564192846417427,
      "num_tokens": 1606313126.0,
      "step": 22350
    },
    {
      "entropy": 0.894322058558464,
      "epoch": 6.094303624965931,
      "grad_norm": 0.1261168122291565,
      "learning_rate": 6.913174284201613e-05,
      "loss": 1.3287,
      "mean_token_accuracy": 0.7501348540186882,
      "num_tokens": 1607023545.0,
      "step": 22360
    },
    {
      "entropy": 0.8885378792881966,
      "epoch": 6.0970291632597435,
      "grad_norm": 0.11853428184986115,
      "learning_rate": 6.910398164488946e-05,
      "loss": 1.3073,
      "mean_token_accuracy": 0.760647115111351,
      "num_tokens": 1607765314.0,
      "step": 22370
    },
    {
      "entropy": 0.8893858775496483,
      "epoch": 6.0997547015535565,
      "grad_norm": 0.12499324977397919,
      "learning_rate": 6.907621355008942e-05,
      "loss": 1.3151,
      "mean_token_accuracy": 0.752248714864254,
      "num_tokens": 1608471248.0,
      "step": 22380
    },
    {
      "entropy": 0.8909812286496163,
      "epoch": 6.102480239847369,
      "grad_norm": 0.1291303187608719,
      "learning_rate": 6.904843856764193e-05,
      "loss": 1.3187,
      "mean_token_accuracy": 0.7530547022819519,
      "num_tokens": 1609188891.0,
      "step": 22390
    },
    {
      "entropy": 0.8938685089349747,
      "epoch": 6.105205778141183,
      "grad_norm": 0.11796817183494568,
      "learning_rate": 6.90206567075754e-05,
      "loss": 1.3156,
      "mean_token_accuracy": 0.7525603711605072,
      "num_tokens": 1609912471.0,
      "step": 22400
    },
    {
      "entropy": 0.8958727091550827,
      "epoch": 6.107931316434996,
      "grad_norm": 0.1305599957704544,
      "learning_rate": 6.899286797992079e-05,
      "loss": 1.3181,
      "mean_token_accuracy": 0.7548779636621475,
      "num_tokens": 1610632297.0,
      "step": 22410
    },
    {
      "entropy": 0.8944883346557617,
      "epoch": 6.110656854728809,
      "grad_norm": 0.12552963197231293,
      "learning_rate": 6.89650723947114e-05,
      "loss": 1.3185,
      "mean_token_accuracy": 0.7562395870685578,
      "num_tokens": 1611369778.0,
      "step": 22420
    },
    {
      "entropy": 0.8903278291225434,
      "epoch": 6.113382393022622,
      "grad_norm": 0.12508676946163177,
      "learning_rate": 6.893726996198315e-05,
      "loss": 1.3099,
      "mean_token_accuracy": 0.7585559323430061,
      "num_tokens": 1612115016.0,
      "step": 22430
    },
    {
      "entropy": 0.8874625027179718,
      "epoch": 6.116107931316435,
      "grad_norm": 0.1288505643606186,
      "learning_rate": 6.890946069177435e-05,
      "loss": 1.3163,
      "mean_token_accuracy": 0.7557416692376137,
      "num_tokens": 1612844894.0,
      "step": 22440
    },
    {
      "entropy": 0.8882686629891395,
      "epoch": 6.118833469610248,
      "grad_norm": 0.12406323850154877,
      "learning_rate": 6.88816445941258e-05,
      "loss": 1.3064,
      "mean_token_accuracy": 0.7553786844015121,
      "num_tokens": 1613566824.0,
      "step": 22450
    },
    {
      "entropy": 0.8838739201426506,
      "epoch": 6.121559007904061,
      "grad_norm": 0.1409529149532318,
      "learning_rate": 6.885382167908077e-05,
      "loss": 1.3128,
      "mean_token_accuracy": 0.7541501834988594,
      "num_tokens": 1614282411.0,
      "step": 22460
    },
    {
      "entropy": 0.895191416144371,
      "epoch": 6.124284546197874,
      "grad_norm": 0.12471000105142593,
      "learning_rate": 6.8825991956685e-05,
      "loss": 1.3214,
      "mean_token_accuracy": 0.754830002784729,
      "num_tokens": 1615002230.0,
      "step": 22470
    },
    {
      "entropy": 0.8889313638210297,
      "epoch": 6.127010084491687,
      "grad_norm": 0.12469696998596191,
      "learning_rate": 6.879815543698665e-05,
      "loss": 1.3122,
      "mean_token_accuracy": 0.7525438651442528,
      "num_tokens": 1615720889.0,
      "step": 22480
    },
    {
      "entropy": 0.8920111954212189,
      "epoch": 6.1297356227855,
      "grad_norm": 0.1284060776233673,
      "learning_rate": 6.877031213003636e-05,
      "loss": 1.321,
      "mean_token_accuracy": 0.7575472369790077,
      "num_tokens": 1616456158.0,
      "step": 22490
    },
    {
      "entropy": 0.8941321104764939,
      "epoch": 6.132461161079313,
      "grad_norm": 0.11869368702173233,
      "learning_rate": 6.874246204588724e-05,
      "loss": 1.309,
      "mean_token_accuracy": 0.7604214429855347,
      "num_tokens": 1617209190.0,
      "step": 22500
    },
    {
      "entropy": 0.8966589763760566,
      "epoch": 6.1351866993731266,
      "grad_norm": 0.1317199319601059,
      "learning_rate": 6.87146051945948e-05,
      "loss": 1.3273,
      "mean_token_accuracy": 0.7514427572488784,
      "num_tokens": 1617929382.0,
      "step": 22510
    },
    {
      "entropy": 0.8902032285928726,
      "epoch": 6.1379122376669395,
      "grad_norm": 0.13315516710281372,
      "learning_rate": 6.868674158621704e-05,
      "loss": 1.3204,
      "mean_token_accuracy": 0.753947240114212,
      "num_tokens": 1618649219.0,
      "step": 22520
    },
    {
      "entropy": 0.8954225808382035,
      "epoch": 6.140637775960752,
      "grad_norm": 0.13436183333396912,
      "learning_rate": 6.865887123081439e-05,
      "loss": 1.3166,
      "mean_token_accuracy": 0.7565888911485672,
      "num_tokens": 1619389312.0,
      "step": 22530
    },
    {
      "entropy": 0.8795060738921165,
      "epoch": 6.143363314254565,
      "grad_norm": 0.13542599976062775,
      "learning_rate": 6.863099413844966e-05,
      "loss": 1.3064,
      "mean_token_accuracy": 0.7552435308694839,
      "num_tokens": 1620104826.0,
      "step": 22540
    },
    {
      "entropy": 0.8968888089060784,
      "epoch": 6.146088852548378,
      "grad_norm": 0.12925991415977478,
      "learning_rate": 6.860311031918819e-05,
      "loss": 1.311,
      "mean_token_accuracy": 0.7561874642968178,
      "num_tokens": 1620847714.0,
      "step": 22550
    },
    {
      "entropy": 0.8927373275160789,
      "epoch": 6.148814390842191,
      "grad_norm": 0.12346458435058594,
      "learning_rate": 6.857521978309767e-05,
      "loss": 1.3223,
      "mean_token_accuracy": 0.7570818364620209,
      "num_tokens": 1621582588.0,
      "step": 22560
    },
    {
      "entropy": 0.8965425729751587,
      "epoch": 6.151539929136004,
      "grad_norm": 0.1410234123468399,
      "learning_rate": 6.854732254024823e-05,
      "loss": 1.3303,
      "mean_token_accuracy": 0.7513458266854286,
      "num_tokens": 1622294199.0,
      "step": 22570
    },
    {
      "entropy": 0.8883983790874481,
      "epoch": 6.154265467429817,
      "grad_norm": 0.1301908940076828,
      "learning_rate": 6.851941860071247e-05,
      "loss": 1.3137,
      "mean_token_accuracy": 0.7528089091181756,
      "num_tokens": 1623003865.0,
      "step": 22580
    },
    {
      "entropy": 0.8913798347115517,
      "epoch": 6.15699100572363,
      "grad_norm": 0.12132018804550171,
      "learning_rate": 6.849150797456536e-05,
      "loss": 1.3174,
      "mean_token_accuracy": 0.7533126696944237,
      "num_tokens": 1623734066.0,
      "step": 22590
    },
    {
      "entropy": 0.8933842197060585,
      "epoch": 6.159716544017443,
      "grad_norm": 0.12742134928703308,
      "learning_rate": 6.846359067188427e-05,
      "loss": 1.3193,
      "mean_token_accuracy": 0.7530615836381912,
      "num_tokens": 1624455959.0,
      "step": 22600
    },
    {
      "entropy": 0.8866404846310616,
      "epoch": 6.162442082311257,
      "grad_norm": 0.13009659945964813,
      "learning_rate": 6.843566670274903e-05,
      "loss": 1.3178,
      "mean_token_accuracy": 0.7511741638183593,
      "num_tokens": 1625156696.0,
      "step": 22610
    },
    {
      "entropy": 0.8912604719400405,
      "epoch": 6.16516762060507,
      "grad_norm": 0.1300330013036728,
      "learning_rate": 6.840773607724184e-05,
      "loss": 1.3175,
      "mean_token_accuracy": 0.7565377578139305,
      "num_tokens": 1625889918.0,
      "step": 22620
    },
    {
      "entropy": 0.8939005225896836,
      "epoch": 6.167893158898883,
      "grad_norm": 0.12857604026794434,
      "learning_rate": 6.837979880544734e-05,
      "loss": 1.322,
      "mean_token_accuracy": 0.7528885453939438,
      "num_tokens": 1626613661.0,
      "step": 22630
    },
    {
      "entropy": 0.8946500390768051,
      "epoch": 6.170618697192696,
      "grad_norm": 0.12966182827949524,
      "learning_rate": 6.835185489745251e-05,
      "loss": 1.3176,
      "mean_token_accuracy": 0.755151455104351,
      "num_tokens": 1627336842.0,
      "step": 22640
    },
    {
      "entropy": 0.8905218884348869,
      "epoch": 6.173344235486509,
      "grad_norm": 0.13915207982063293,
      "learning_rate": 6.832390436334679e-05,
      "loss": 1.3141,
      "mean_token_accuracy": 0.7541632756590844,
      "num_tokens": 1628056044.0,
      "step": 22650
    },
    {
      "entropy": 0.888375709950924,
      "epoch": 6.176069773780322,
      "grad_norm": 0.1322900652885437,
      "learning_rate": 6.829594721322198e-05,
      "loss": 1.3194,
      "mean_token_accuracy": 0.7544383108615875,
      "num_tokens": 1628772990.0,
      "step": 22660
    },
    {
      "entropy": 0.8861851319670677,
      "epoch": 6.178795312074135,
      "grad_norm": 0.12365427613258362,
      "learning_rate": 6.826798345717227e-05,
      "loss": 1.3154,
      "mean_token_accuracy": 0.7540062472224236,
      "num_tokens": 1629491637.0,
      "step": 22670
    },
    {
      "entropy": 0.8840339452028274,
      "epoch": 6.1815208503679475,
      "grad_norm": 0.12385515123605728,
      "learning_rate": 6.824001310529424e-05,
      "loss": 1.3083,
      "mean_token_accuracy": 0.7553656801581383,
      "num_tokens": 1630209301.0,
      "step": 22680
    },
    {
      "entropy": 0.904748423397541,
      "epoch": 6.1842463886617605,
      "grad_norm": 0.11822356283664703,
      "learning_rate": 6.821203616768685e-05,
      "loss": 1.3172,
      "mean_token_accuracy": 0.7522397011518478,
      "num_tokens": 1630942852.0,
      "step": 22690
    },
    {
      "entropy": 0.8997132733464241,
      "epoch": 6.186971926955573,
      "grad_norm": 0.13043208420276642,
      "learning_rate": 6.818405265445143e-05,
      "loss": 1.3225,
      "mean_token_accuracy": 0.7487904891371727,
      "num_tokens": 1631646181.0,
      "step": 22700
    },
    {
      "entropy": 0.8879512473940849,
      "epoch": 6.189697465249386,
      "grad_norm": 0.12371441721916199,
      "learning_rate": 6.815606257569171e-05,
      "loss": 1.3124,
      "mean_token_accuracy": 0.7555837824940681,
      "num_tokens": 1632373271.0,
      "step": 22710
    },
    {
      "entropy": 0.8926724269986153,
      "epoch": 6.1924230035432,
      "grad_norm": 0.1282380223274231,
      "learning_rate": 6.812806594151377e-05,
      "loss": 1.3153,
      "mean_token_accuracy": 0.7546018064022064,
      "num_tokens": 1633103388.0,
      "step": 22720
    },
    {
      "entropy": 0.8907320380210877,
      "epoch": 6.195148541837013,
      "grad_norm": 0.12031175941228867,
      "learning_rate": 6.810006276202603e-05,
      "loss": 1.3119,
      "mean_token_accuracy": 0.7557961121201515,
      "num_tokens": 1633832517.0,
      "step": 22730
    },
    {
      "entropy": 0.8913863524794579,
      "epoch": 6.197874080130826,
      "grad_norm": 0.12618345022201538,
      "learning_rate": 6.807205304733936e-05,
      "loss": 1.3171,
      "mean_token_accuracy": 0.7520734235644341,
      "num_tokens": 1634544327.0,
      "step": 22740
    },
    {
      "entropy": 0.8889236986637116,
      "epoch": 6.200599618424639,
      "grad_norm": 0.1265101432800293,
      "learning_rate": 6.804403680756687e-05,
      "loss": 1.3143,
      "mean_token_accuracy": 0.755327270925045,
      "num_tokens": 1635271233.0,
      "step": 22750
    },
    {
      "entropy": 0.8882743090391159,
      "epoch": 6.203325156718452,
      "grad_norm": 0.12307625263929367,
      "learning_rate": 6.801601405282412e-05,
      "loss": 1.3173,
      "mean_token_accuracy": 0.7541785255074501,
      "num_tokens": 1635991568.0,
      "step": 22760
    },
    {
      "entropy": 0.8878097742795944,
      "epoch": 6.206050695012265,
      "grad_norm": 0.12814800441265106,
      "learning_rate": 6.7987984793229e-05,
      "loss": 1.3119,
      "mean_token_accuracy": 0.7537838399410248,
      "num_tokens": 1636697369.0,
      "step": 22770
    },
    {
      "entropy": 0.8963440015912056,
      "epoch": 6.208776233306078,
      "grad_norm": 0.1348268836736679,
      "learning_rate": 6.795994903890172e-05,
      "loss": 1.318,
      "mean_token_accuracy": 0.7527847737073898,
      "num_tokens": 1637423328.0,
      "step": 22780
    },
    {
      "entropy": 0.8971793323755264,
      "epoch": 6.211501771599891,
      "grad_norm": 0.12599802017211914,
      "learning_rate": 6.793190679996487e-05,
      "loss": 1.3183,
      "mean_token_accuracy": 0.7480891391634941,
      "num_tokens": 1638136778.0,
      "step": 22790
    },
    {
      "entropy": 0.8978354439139367,
      "epoch": 6.214227309893704,
      "grad_norm": 0.11951422691345215,
      "learning_rate": 6.790385808654334e-05,
      "loss": 1.329,
      "mean_token_accuracy": 0.7470620110630989,
      "num_tokens": 1638845767.0,
      "step": 22800
    },
    {
      "entropy": 0.8890793532133102,
      "epoch": 6.216952848187517,
      "grad_norm": 0.14458394050598145,
      "learning_rate": 6.787580290876441e-05,
      "loss": 1.3214,
      "mean_token_accuracy": 0.7524440795183182,
      "num_tokens": 1639554711.0,
      "step": 22810
    },
    {
      "entropy": 0.8881776869297028,
      "epoch": 6.21967838648133,
      "grad_norm": 0.1261187046766281,
      "learning_rate": 6.784774127675766e-05,
      "loss": 1.3093,
      "mean_token_accuracy": 0.7549481615424156,
      "num_tokens": 1640285285.0,
      "step": 22820
    },
    {
      "entropy": 0.8906338453292847,
      "epoch": 6.2224039247751435,
      "grad_norm": 0.1307862251996994,
      "learning_rate": 6.781967320065501e-05,
      "loss": 1.3079,
      "mean_token_accuracy": 0.7568583205342293,
      "num_tokens": 1641016695.0,
      "step": 22830
    },
    {
      "entropy": 0.8807773485779762,
      "epoch": 6.225129463068956,
      "grad_norm": 0.1270526498556137,
      "learning_rate": 6.77915986905907e-05,
      "loss": 1.3095,
      "mean_token_accuracy": 0.7582296550273895,
      "num_tokens": 1641747377.0,
      "step": 22840
    },
    {
      "entropy": 0.8841337040066719,
      "epoch": 6.227855001362769,
      "grad_norm": 0.12727545201778412,
      "learning_rate": 6.776351775670129e-05,
      "loss": 1.3074,
      "mean_token_accuracy": 0.7586832597851754,
      "num_tokens": 1642474278.0,
      "step": 22850
    },
    {
      "entropy": 0.8928334191441536,
      "epoch": 6.230580539656582,
      "grad_norm": 0.12225507199764252,
      "learning_rate": 6.773543040912569e-05,
      "loss": 1.321,
      "mean_token_accuracy": 0.7530438497662544,
      "num_tokens": 1643185876.0,
      "step": 22860
    },
    {
      "entropy": 0.895265993475914,
      "epoch": 6.233306077950395,
      "grad_norm": 0.1776934713125229,
      "learning_rate": 6.770733665800507e-05,
      "loss": 1.3236,
      "mean_token_accuracy": 0.7498766735196114,
      "num_tokens": 1643890972.0,
      "step": 22870
    },
    {
      "entropy": 0.8864861071109772,
      "epoch": 6.236031616244208,
      "grad_norm": 0.12553735077381134,
      "learning_rate": 6.767923651348297e-05,
      "loss": 1.3186,
      "mean_token_accuracy": 0.7555804580450058,
      "num_tokens": 1644600926.0,
      "step": 22880
    },
    {
      "entropy": 0.8897723659873009,
      "epoch": 6.238757154538021,
      "grad_norm": 0.12780770659446716,
      "learning_rate": 6.765112998570522e-05,
      "loss": 1.3134,
      "mean_token_accuracy": 0.7504253312945366,
      "num_tokens": 1645309227.0,
      "step": 22890
    },
    {
      "entropy": 0.8836939483880997,
      "epoch": 6.241482692831834,
      "grad_norm": 0.13114844262599945,
      "learning_rate": 6.76230170848199e-05,
      "loss": 1.3151,
      "mean_token_accuracy": 0.7565981298685074,
      "num_tokens": 1646027636.0,
      "step": 22900
    },
    {
      "entropy": 0.8949156388640404,
      "epoch": 6.244208231125647,
      "grad_norm": 0.12638813257217407,
      "learning_rate": 6.759489782097749e-05,
      "loss": 1.3168,
      "mean_token_accuracy": 0.7493349418044091,
      "num_tokens": 1646748504.0,
      "step": 22910
    },
    {
      "entropy": 0.8914054661989212,
      "epoch": 6.24693376941946,
      "grad_norm": 0.12566597759723663,
      "learning_rate": 6.756677220433072e-05,
      "loss": 1.3223,
      "mean_token_accuracy": 0.7530945912003517,
      "num_tokens": 1647465546.0,
      "step": 22920
    },
    {
      "entropy": 0.8930309310555458,
      "epoch": 6.249659307713273,
      "grad_norm": 0.1308824121952057,
      "learning_rate": 6.753864024503458e-05,
      "loss": 1.323,
      "mean_token_accuracy": 0.7517623871564865,
      "num_tokens": 1648180371.0,
      "step": 22930
    },
    {
      "entropy": 0.8952459499239922,
      "epoch": 6.252384846007087,
      "grad_norm": 0.13122569024562836,
      "learning_rate": 6.751050195324638e-05,
      "loss": 1.3199,
      "mean_token_accuracy": 0.749671746790409,
      "num_tokens": 1648882516.0,
      "step": 22940
    },
    {
      "entropy": 0.8864050194621086,
      "epoch": 6.2551103843009,
      "grad_norm": 0.1290164291858673,
      "learning_rate": 6.748235733912573e-05,
      "loss": 1.3116,
      "mean_token_accuracy": 0.7586789190769195,
      "num_tokens": 1649608654.0,
      "step": 22950
    },
    {
      "entropy": 0.8909218072891235,
      "epoch": 6.257835922594713,
      "grad_norm": 0.12734967470169067,
      "learning_rate": 6.745420641283449e-05,
      "loss": 1.3201,
      "mean_token_accuracy": 0.7540360197424889,
      "num_tokens": 1650323619.0,
      "step": 22960
    },
    {
      "entropy": 0.894496864080429,
      "epoch": 6.260561460888526,
      "grad_norm": 0.11867977678775787,
      "learning_rate": 6.742604918453684e-05,
      "loss": 1.3137,
      "mean_token_accuracy": 0.7555643871426583,
      "num_tokens": 1651061365.0,
      "step": 22970
    },
    {
      "entropy": 0.8806008353829384,
      "epoch": 6.2632869991823386,
      "grad_norm": 0.1332656443119049,
      "learning_rate": 6.739788566439923e-05,
      "loss": 1.3198,
      "mean_token_accuracy": 0.7565188840031624,
      "num_tokens": 1651776100.0,
      "step": 22980
    },
    {
      "entropy": 0.8906694203615189,
      "epoch": 6.2660125374761515,
      "grad_norm": 0.1220070943236351,
      "learning_rate": 6.736971586259033e-05,
      "loss": 1.3216,
      "mean_token_accuracy": 0.7555127948522568,
      "num_tokens": 1652500151.0,
      "step": 22990
    },
    {
      "entropy": 0.8953827202320099,
      "epoch": 6.2687380757699644,
      "grad_norm": 0.1308388113975525,
      "learning_rate": 6.734153978928111e-05,
      "loss": 1.3225,
      "mean_token_accuracy": 0.7522341355681419,
      "num_tokens": 1653217232.0,
      "step": 23000
    },
    {
      "entropy": 0.8931818947196006,
      "epoch": 6.271463614063777,
      "grad_norm": 0.13166120648384094,
      "learning_rate": 6.731335745464485e-05,
      "loss": 1.3225,
      "mean_token_accuracy": 0.7508315578103065,
      "num_tokens": 1653926371.0,
      "step": 23010
    },
    {
      "entropy": 0.8927141845226287,
      "epoch": 6.27418915235759,
      "grad_norm": 0.12071728706359863,
      "learning_rate": 6.7285168868857e-05,
      "loss": 1.3139,
      "mean_token_accuracy": 0.7525391235947609,
      "num_tokens": 1654653506.0,
      "step": 23020
    },
    {
      "entropy": 0.8891505330801011,
      "epoch": 6.276914690651403,
      "grad_norm": 0.13824611902236938,
      "learning_rate": 6.725697404209535e-05,
      "loss": 1.3117,
      "mean_token_accuracy": 0.7533309891819954,
      "num_tokens": 1655366801.0,
      "step": 23030
    },
    {
      "entropy": 0.8925409883260726,
      "epoch": 6.279640228945217,
      "grad_norm": 0.12709136307239532,
      "learning_rate": 6.722877298453988e-05,
      "loss": 1.3151,
      "mean_token_accuracy": 0.7516300871968269,
      "num_tokens": 1656084379.0,
      "step": 23040
    },
    {
      "entropy": 0.8974668428301811,
      "epoch": 6.28236576723903,
      "grad_norm": 0.135416641831398,
      "learning_rate": 6.720056570637288e-05,
      "loss": 1.3193,
      "mean_token_accuracy": 0.7500672295689583,
      "num_tokens": 1656795900.0,
      "step": 23050
    },
    {
      "entropy": 0.8910537645220756,
      "epoch": 6.285091305532843,
      "grad_norm": 0.12810403108596802,
      "learning_rate": 6.717235221777884e-05,
      "loss": 1.3158,
      "mean_token_accuracy": 0.7524344757199287,
      "num_tokens": 1657517916.0,
      "step": 23060
    },
    {
      "entropy": 0.895925197005272,
      "epoch": 6.287816843826656,
      "grad_norm": 0.12679268419742584,
      "learning_rate": 6.714413252894452e-05,
      "loss": 1.3199,
      "mean_token_accuracy": 0.7524157360196113,
      "num_tokens": 1658236479.0,
      "step": 23070
    },
    {
      "entropy": 0.8970654159784317,
      "epoch": 6.290542382120469,
      "grad_norm": 0.12934447824954987,
      "learning_rate": 6.711590665005887e-05,
      "loss": 1.3296,
      "mean_token_accuracy": 0.7537188276648521,
      "num_tokens": 1658958245.0,
      "step": 23080
    },
    {
      "entropy": 0.8944193720817566,
      "epoch": 6.293267920414282,
      "grad_norm": 0.12345714122056961,
      "learning_rate": 6.708767459131315e-05,
      "loss": 1.3269,
      "mean_token_accuracy": 0.7518907606601715,
      "num_tokens": 1659665312.0,
      "step": 23090
    },
    {
      "entropy": 0.8892689511179924,
      "epoch": 6.295993458708095,
      "grad_norm": 0.1303255259990692,
      "learning_rate": 6.705943636290079e-05,
      "loss": 1.3227,
      "mean_token_accuracy": 0.7500828951597214,
      "num_tokens": 1660368213.0,
      "step": 23100
    },
    {
      "entropy": 0.8863576143980026,
      "epoch": 6.298718997001908,
      "grad_norm": 0.12337734550237656,
      "learning_rate": 6.70311919750175e-05,
      "loss": 1.3066,
      "mean_token_accuracy": 0.7597868055105209,
      "num_tokens": 1661104476.0,
      "step": 23110
    },
    {
      "entropy": 0.8880080595612526,
      "epoch": 6.301444535295721,
      "grad_norm": 0.12878258526325226,
      "learning_rate": 6.700294143786115e-05,
      "loss": 1.3135,
      "mean_token_accuracy": 0.7550767511129379,
      "num_tokens": 1661826814.0,
      "step": 23120
    },
    {
      "entropy": 0.8858757704496384,
      "epoch": 6.304170073589534,
      "grad_norm": 0.13227084279060364,
      "learning_rate": 6.697468476163185e-05,
      "loss": 1.314,
      "mean_token_accuracy": 0.7593360260128975,
      "num_tokens": 1662562879.0,
      "step": 23130
    },
    {
      "entropy": 0.8885963514447213,
      "epoch": 6.306895611883347,
      "grad_norm": 0.14592814445495605,
      "learning_rate": 6.694642195653198e-05,
      "loss": 1.324,
      "mean_token_accuracy": 0.7531319305300712,
      "num_tokens": 1663273635.0,
      "step": 23140
    },
    {
      "entropy": 0.8851574152708054,
      "epoch": 6.3096211501771595,
      "grad_norm": 0.13013918697834015,
      "learning_rate": 6.691815303276608e-05,
      "loss": 1.3048,
      "mean_token_accuracy": 0.7523117408156395,
      "num_tokens": 1663991083.0,
      "step": 23150
    },
    {
      "entropy": 0.8889756724238396,
      "epoch": 6.312346688470973,
      "grad_norm": 0.12669463455677032,
      "learning_rate": 6.68898780005409e-05,
      "loss": 1.3156,
      "mean_token_accuracy": 0.7576033070683479,
      "num_tokens": 1664715614.0,
      "step": 23160
    },
    {
      "entropy": 0.8899347677826881,
      "epoch": 6.315072226764786,
      "grad_norm": 0.12626436352729797,
      "learning_rate": 6.686159687006541e-05,
      "loss": 1.3224,
      "mean_token_accuracy": 0.7536514043807984,
      "num_tokens": 1665445966.0,
      "step": 23170
    },
    {
      "entropy": 0.8888762414455413,
      "epoch": 6.317797765058599,
      "grad_norm": 0.12010875344276428,
      "learning_rate": 6.683330965155081e-05,
      "loss": 1.3095,
      "mean_token_accuracy": 0.7526542827486992,
      "num_tokens": 1666160729.0,
      "step": 23180
    },
    {
      "entropy": 0.8972968846559525,
      "epoch": 6.320523303352412,
      "grad_norm": 0.12839588522911072,
      "learning_rate": 6.680501635521043e-05,
      "loss": 1.3226,
      "mean_token_accuracy": 0.7494635313749314,
      "num_tokens": 1666879523.0,
      "step": 23190
    },
    {
      "entropy": 0.9004280254244804,
      "epoch": 6.323248841646225,
      "grad_norm": 0.1343114674091339,
      "learning_rate": 6.677671699125986e-05,
      "loss": 1.3275,
      "mean_token_accuracy": 0.748376627266407,
      "num_tokens": 1667591974.0,
      "step": 23200
    },
    {
      "entropy": 0.9000216856598854,
      "epoch": 6.325974379940038,
      "grad_norm": 0.12477145344018936,
      "learning_rate": 6.674841156991685e-05,
      "loss": 1.3271,
      "mean_token_accuracy": 0.7486219316720962,
      "num_tokens": 1668299821.0,
      "step": 23210
    },
    {
      "entropy": 0.8939940422773361,
      "epoch": 6.328699918233851,
      "grad_norm": 0.1266832947731018,
      "learning_rate": 6.67201001014013e-05,
      "loss": 1.3179,
      "mean_token_accuracy": 0.752446548640728,
      "num_tokens": 1669024776.0,
      "step": 23220
    },
    {
      "entropy": 0.8983014822006226,
      "epoch": 6.331425456527664,
      "grad_norm": 0.1296415477991104,
      "learning_rate": 6.669178259593538e-05,
      "loss": 1.3265,
      "mean_token_accuracy": 0.7508083760738373,
      "num_tokens": 1669750314.0,
      "step": 23230
    },
    {
      "entropy": 0.8913666635751725,
      "epoch": 6.334150994821477,
      "grad_norm": 0.13265082240104675,
      "learning_rate": 6.666345906374338e-05,
      "loss": 1.3269,
      "mean_token_accuracy": 0.7540781378746033,
      "num_tokens": 1670467847.0,
      "step": 23240
    },
    {
      "entropy": 0.8890855371952057,
      "epoch": 6.33687653311529,
      "grad_norm": 0.126531720161438,
      "learning_rate": 6.663512951505177e-05,
      "loss": 1.3054,
      "mean_token_accuracy": 0.7585485383868218,
      "num_tokens": 1671200568.0,
      "step": 23250
    },
    {
      "entropy": 0.9016404092311859,
      "epoch": 6.339602071409104,
      "grad_norm": 0.1255868375301361,
      "learning_rate": 6.660679396008922e-05,
      "loss": 1.3323,
      "mean_token_accuracy": 0.7508169993758201,
      "num_tokens": 1671915657.0,
      "step": 23260
    },
    {
      "entropy": 0.8987109735608101,
      "epoch": 6.342327609702917,
      "grad_norm": 0.14064787328243256,
      "learning_rate": 6.657845240908654e-05,
      "loss": 1.3177,
      "mean_token_accuracy": 0.7515898689627647,
      "num_tokens": 1672622573.0,
      "step": 23270
    },
    {
      "entropy": 0.8917651459574699,
      "epoch": 6.34505314799673,
      "grad_norm": 0.12363500893115997,
      "learning_rate": 6.655010487227671e-05,
      "loss": 1.3151,
      "mean_token_accuracy": 0.7533663526177407,
      "num_tokens": 1673335813.0,
      "step": 23280
    },
    {
      "entropy": 0.8902298092842102,
      "epoch": 6.3477786862905425,
      "grad_norm": 0.13857240974903107,
      "learning_rate": 6.652175135989487e-05,
      "loss": 1.3145,
      "mean_token_accuracy": 0.7535278961062432,
      "num_tokens": 1674046266.0,
      "step": 23290
    },
    {
      "entropy": 0.8870541349053382,
      "epoch": 6.3505042245843555,
      "grad_norm": 0.15220853686332703,
      "learning_rate": 6.649339188217836e-05,
      "loss": 1.3048,
      "mean_token_accuracy": 0.7575495153665542,
      "num_tokens": 1674767492.0,
      "step": 23300
    },
    {
      "entropy": 0.8958998307585716,
      "epoch": 6.353229762878168,
      "grad_norm": 0.12492650747299194,
      "learning_rate": 6.646502644936659e-05,
      "loss": 1.3169,
      "mean_token_accuracy": 0.7527160704135895,
      "num_tokens": 1675501332.0,
      "step": 23310
    },
    {
      "entropy": 0.8936477765440941,
      "epoch": 6.355955301171981,
      "grad_norm": 0.13287654519081116,
      "learning_rate": 6.64366550717012e-05,
      "loss": 1.313,
      "mean_token_accuracy": 0.7562225490808487,
      "num_tokens": 1676238175.0,
      "step": 23320
    },
    {
      "entropy": 0.8891023188829422,
      "epoch": 6.358680839465794,
      "grad_norm": 0.13067872822284698,
      "learning_rate": 6.640827775942595e-05,
      "loss": 1.3143,
      "mean_token_accuracy": 0.7531682774424553,
      "num_tokens": 1676959889.0,
      "step": 23330
    },
    {
      "entropy": 0.8914091661572456,
      "epoch": 6.361406377759607,
      "grad_norm": 0.1234472244977951,
      "learning_rate": 6.63798945227867e-05,
      "loss": 1.3227,
      "mean_token_accuracy": 0.7489127948880195,
      "num_tokens": 1677665496.0,
      "step": 23340
    },
    {
      "entropy": 0.8856423974037171,
      "epoch": 6.36413191605342,
      "grad_norm": 0.13285700976848602,
      "learning_rate": 6.635150537203152e-05,
      "loss": 1.3144,
      "mean_token_accuracy": 0.7498030632734298,
      "num_tokens": 1678366349.0,
      "step": 23350
    },
    {
      "entropy": 0.8926140740513802,
      "epoch": 6.366857454347233,
      "grad_norm": 0.12738476693630219,
      "learning_rate": 6.632311031741057e-05,
      "loss": 1.3163,
      "mean_token_accuracy": 0.7527148187160492,
      "num_tokens": 1679081477.0,
      "step": 23360
    },
    {
      "entropy": 0.8965121746063233,
      "epoch": 6.369582992641047,
      "grad_norm": 0.12689495086669922,
      "learning_rate": 6.629470936917615e-05,
      "loss": 1.3299,
      "mean_token_accuracy": 0.7492060095071793,
      "num_tokens": 1679790276.0,
      "step": 23370
    },
    {
      "entropy": 0.8893764913082123,
      "epoch": 6.37230853093486,
      "grad_norm": 0.12410731613636017,
      "learning_rate": 6.62663025375827e-05,
      "loss": 1.3098,
      "mean_token_accuracy": 0.74893439412117,
      "num_tokens": 1680495360.0,
      "step": 23380
    },
    {
      "entropy": 0.8907683804631233,
      "epoch": 6.375034069228673,
      "grad_norm": 0.13427311182022095,
      "learning_rate": 6.623788983288677e-05,
      "loss": 1.3198,
      "mean_token_accuracy": 0.754366135597229,
      "num_tokens": 1681219401.0,
      "step": 23390
    },
    {
      "entropy": 0.896857637166977,
      "epoch": 6.377759607522486,
      "grad_norm": 0.12843424081802368,
      "learning_rate": 6.620947126534703e-05,
      "loss": 1.3208,
      "mean_token_accuracy": 0.7514572486281395,
      "num_tokens": 1681936126.0,
      "step": 23400
    },
    {
      "entropy": 0.8925153493881226,
      "epoch": 6.380485145816299,
      "grad_norm": 0.1244444027543068,
      "learning_rate": 6.618104684522428e-05,
      "loss": 1.3274,
      "mean_token_accuracy": 0.7482243925333023,
      "num_tokens": 1682636547.0,
      "step": 23410
    },
    {
      "entropy": 0.8861335068941116,
      "epoch": 6.383210684110112,
      "grad_norm": 0.142676442861557,
      "learning_rate": 6.615261658278144e-05,
      "loss": 1.3196,
      "mean_token_accuracy": 0.7547627225518226,
      "num_tokens": 1683340915.0,
      "step": 23420
    },
    {
      "entropy": 0.8925488859415054,
      "epoch": 6.385936222403925,
      "grad_norm": 0.12367556244134903,
      "learning_rate": 6.612418048828349e-05,
      "loss": 1.3169,
      "mean_token_accuracy": 0.7496591046452522,
      "num_tokens": 1684046546.0,
      "step": 23430
    },
    {
      "entropy": 0.8861442267894745,
      "epoch": 6.388661760697738,
      "grad_norm": 0.13329732418060303,
      "learning_rate": 6.609573857199758e-05,
      "loss": 1.3137,
      "mean_token_accuracy": 0.753875145316124,
      "num_tokens": 1684753967.0,
      "step": 23440
    },
    {
      "entropy": 0.8948607563972473,
      "epoch": 6.391387298991551,
      "grad_norm": 0.13758523762226105,
      "learning_rate": 6.606729084419293e-05,
      "loss": 1.3259,
      "mean_token_accuracy": 0.7501895040273666,
      "num_tokens": 1685458624.0,
      "step": 23450
    },
    {
      "entropy": 0.8905760660767555,
      "epoch": 6.3941128372853635,
      "grad_norm": 0.13088218867778778,
      "learning_rate": 6.603883731514084e-05,
      "loss": 1.3139,
      "mean_token_accuracy": 0.7591215088963509,
      "num_tokens": 1686191114.0,
      "step": 23460
    },
    {
      "entropy": 0.8919359445571899,
      "epoch": 6.396838375579177,
      "grad_norm": 0.13223186135292053,
      "learning_rate": 6.601037799511474e-05,
      "loss": 1.3256,
      "mean_token_accuracy": 0.7524249017238617,
      "num_tokens": 1686898721.0,
      "step": 23470
    },
    {
      "entropy": 0.8904221951961517,
      "epoch": 6.39956391387299,
      "grad_norm": 0.12571395933628082,
      "learning_rate": 6.598191289439016e-05,
      "loss": 1.3139,
      "mean_token_accuracy": 0.7535535246133804,
      "num_tokens": 1687611162.0,
      "step": 23480
    },
    {
      "entropy": 0.8901495292782784,
      "epoch": 6.402289452166803,
      "grad_norm": 0.13453614711761475,
      "learning_rate": 6.595344202324468e-05,
      "loss": 1.3244,
      "mean_token_accuracy": 0.7533126056194306,
      "num_tokens": 1688325875.0,
      "step": 23490
    },
    {
      "entropy": 0.8958468347787857,
      "epoch": 6.405014990460616,
      "grad_norm": 0.1318880021572113,
      "learning_rate": 6.592496539195794e-05,
      "loss": 1.325,
      "mean_token_accuracy": 0.7493641376495361,
      "num_tokens": 1689030804.0,
      "step": 23500
    },
    {
      "entropy": 0.8912514030933381,
      "epoch": 6.407740528754429,
      "grad_norm": 0.13106226921081543,
      "learning_rate": 6.589648301081173e-05,
      "loss": 1.3185,
      "mean_token_accuracy": 0.7552205622196198,
      "num_tokens": 1689761757.0,
      "step": 23510
    },
    {
      "entropy": 0.8896019339561463,
      "epoch": 6.410466067048242,
      "grad_norm": 0.1246907114982605,
      "learning_rate": 6.586799489008988e-05,
      "loss": 1.3131,
      "mean_token_accuracy": 0.7572718515992165,
      "num_tokens": 1690498781.0,
      "step": 23520
    },
    {
      "entropy": 0.8915984958410264,
      "epoch": 6.413191605342055,
      "grad_norm": 0.13249586522579193,
      "learning_rate": 6.58395010400783e-05,
      "loss": 1.312,
      "mean_token_accuracy": 0.7548464789986611,
      "num_tokens": 1691226504.0,
      "step": 23530
    },
    {
      "entropy": 0.8985194623470306,
      "epoch": 6.415917143635868,
      "grad_norm": 0.12803570926189423,
      "learning_rate": 6.581100147106498e-05,
      "loss": 1.3178,
      "mean_token_accuracy": 0.755246813595295,
      "num_tokens": 1691964479.0,
      "step": 23540
    },
    {
      "entropy": 0.8917974069714546,
      "epoch": 6.418642681929681,
      "grad_norm": 0.13058391213417053,
      "learning_rate": 6.578249619333992e-05,
      "loss": 1.3173,
      "mean_token_accuracy": 0.7533485591411591,
      "num_tokens": 1692680451.0,
      "step": 23550
    },
    {
      "entropy": 0.889213714003563,
      "epoch": 6.421368220223494,
      "grad_norm": 0.1276865452528,
      "learning_rate": 6.575398521719525e-05,
      "loss": 1.3171,
      "mean_token_accuracy": 0.753696845471859,
      "num_tokens": 1693391615.0,
      "step": 23560
    },
    {
      "entropy": 0.8873524084687233,
      "epoch": 6.424093758517307,
      "grad_norm": 0.1355648934841156,
      "learning_rate": 6.572546855292511e-05,
      "loss": 1.3178,
      "mean_token_accuracy": 0.7564250737428665,
      "num_tokens": 1694103861.0,
      "step": 23570
    },
    {
      "entropy": 0.889432792365551,
      "epoch": 6.42681929681112,
      "grad_norm": 0.1281968355178833,
      "learning_rate": 6.569694621082572e-05,
      "loss": 1.3159,
      "mean_token_accuracy": 0.7529659688472747,
      "num_tokens": 1694821987.0,
      "step": 23580
    },
    {
      "entropy": 0.8942313596606255,
      "epoch": 6.429544835104934,
      "grad_norm": 0.13217511773109436,
      "learning_rate": 6.566841820119534e-05,
      "loss": 1.3156,
      "mean_token_accuracy": 0.7584445595741272,
      "num_tokens": 1695555333.0,
      "step": 23590
    },
    {
      "entropy": 0.8878136515617371,
      "epoch": 6.4322703733987465,
      "grad_norm": 0.15396036207675934,
      "learning_rate": 6.563988453433429e-05,
      "loss": 1.3119,
      "mean_token_accuracy": 0.7593174755573273,
      "num_tokens": 1696294307.0,
      "step": 23600
    },
    {
      "entropy": 0.8902526319026947,
      "epoch": 6.4349959116925595,
      "grad_norm": 0.12914952635765076,
      "learning_rate": 6.561134522054488e-05,
      "loss": 1.3163,
      "mean_token_accuracy": 0.7554532706737518,
      "num_tokens": 1697025205.0,
      "step": 23610
    },
    {
      "entropy": 0.8902714803814888,
      "epoch": 6.437721449986372,
      "grad_norm": 0.12394432723522186,
      "learning_rate": 6.558280027013154e-05,
      "loss": 1.3202,
      "mean_token_accuracy": 0.7533359780907631,
      "num_tokens": 1697746596.0,
      "step": 23620
    },
    {
      "entropy": 0.9008207753300667,
      "epoch": 6.440446988280185,
      "grad_norm": 0.11941054463386536,
      "learning_rate": 6.555424969340071e-05,
      "loss": 1.3303,
      "mean_token_accuracy": 0.7471015900373459,
      "num_tokens": 1698454352.0,
      "step": 23630
    },
    {
      "entropy": 0.8937665760517121,
      "epoch": 6.443172526573998,
      "grad_norm": 0.1217990517616272,
      "learning_rate": 6.55256935006608e-05,
      "loss": 1.3055,
      "mean_token_accuracy": 0.750526025891304,
      "num_tokens": 1699170833.0,
      "step": 23640
    },
    {
      "entropy": 0.8947145149111748,
      "epoch": 6.445898064867811,
      "grad_norm": 0.17192257940769196,
      "learning_rate": 6.54971317022223e-05,
      "loss": 1.3328,
      "mean_token_accuracy": 0.7513601586222649,
      "num_tokens": 1699883942.0,
      "step": 23650
    },
    {
      "entropy": 0.9000633776187896,
      "epoch": 6.448623603161624,
      "grad_norm": 0.12901583313941956,
      "learning_rate": 6.546856430839776e-05,
      "loss": 1.3234,
      "mean_token_accuracy": 0.7515505746006965,
      "num_tokens": 1700602474.0,
      "step": 23660
    },
    {
      "entropy": 0.8940084531903267,
      "epoch": 6.451349141455437,
      "grad_norm": 0.14255863428115845,
      "learning_rate": 6.543999132950165e-05,
      "loss": 1.3258,
      "mean_token_accuracy": 0.7506169125437736,
      "num_tokens": 1701309757.0,
      "step": 23670
    },
    {
      "entropy": 0.8904284313321114,
      "epoch": 6.45407467974925,
      "grad_norm": 0.14267265796661377,
      "learning_rate": 6.541141277585054e-05,
      "loss": 1.3157,
      "mean_token_accuracy": 0.754489229619503,
      "num_tokens": 1702031621.0,
      "step": 23680
    },
    {
      "entropy": 0.8959540873765945,
      "epoch": 6.456800218043064,
      "grad_norm": 0.13613547384738922,
      "learning_rate": 6.5382828657763e-05,
      "loss": 1.3227,
      "mean_token_accuracy": 0.75272366553545,
      "num_tokens": 1702753239.0,
      "step": 23690
    },
    {
      "entropy": 0.8928972169756889,
      "epoch": 6.459525756336877,
      "grad_norm": 0.12474264949560165,
      "learning_rate": 6.535423898555958e-05,
      "loss": 1.3129,
      "mean_token_accuracy": 0.754960723221302,
      "num_tokens": 1703474389.0,
      "step": 23700
    },
    {
      "entropy": 0.8876286342740058,
      "epoch": 6.46225129463069,
      "grad_norm": 0.12610182166099548,
      "learning_rate": 6.532564376956289e-05,
      "loss": 1.3154,
      "mean_token_accuracy": 0.7527513980865479,
      "num_tokens": 1704188643.0,
      "step": 23710
    },
    {
      "entropy": 0.8942540109157562,
      "epoch": 6.464976832924503,
      "grad_norm": 0.13448436558246613,
      "learning_rate": 6.529704302009743e-05,
      "loss": 1.3208,
      "mean_token_accuracy": 0.7528475806117058,
      "num_tokens": 1704915281.0,
      "step": 23720
    },
    {
      "entropy": 0.892904594540596,
      "epoch": 6.467702371218316,
      "grad_norm": 0.12359185516834259,
      "learning_rate": 6.526843674748983e-05,
      "loss": 1.3204,
      "mean_token_accuracy": 0.7529234677553177,
      "num_tokens": 1705641134.0,
      "step": 23730
    },
    {
      "entropy": 0.895223182439804,
      "epoch": 6.470427909512129,
      "grad_norm": 0.13306471705436707,
      "learning_rate": 6.523982496206863e-05,
      "loss": 1.3143,
      "mean_token_accuracy": 0.751647324860096,
      "num_tokens": 1706364710.0,
      "step": 23740
    },
    {
      "entropy": 0.8945942804217338,
      "epoch": 6.473153447805942,
      "grad_norm": 0.13376927375793457,
      "learning_rate": 6.52112076741644e-05,
      "loss": 1.3173,
      "mean_token_accuracy": 0.7524019733071328,
      "num_tokens": 1707091616.0,
      "step": 23750
    },
    {
      "entropy": 0.8862835958600044,
      "epoch": 6.4758789860997545,
      "grad_norm": 0.13562124967575073,
      "learning_rate": 6.518258489410967e-05,
      "loss": 1.3113,
      "mean_token_accuracy": 0.757615952193737,
      "num_tokens": 1707810977.0,
      "step": 23760
    },
    {
      "entropy": 0.8875472873449326,
      "epoch": 6.4786045243935675,
      "grad_norm": 0.12648871541023254,
      "learning_rate": 6.515395663223898e-05,
      "loss": 1.3166,
      "mean_token_accuracy": 0.7528648927807808,
      "num_tokens": 1708514829.0,
      "step": 23770
    },
    {
      "entropy": 0.8876741334795952,
      "epoch": 6.48133006268738,
      "grad_norm": 0.1349627673625946,
      "learning_rate": 6.512532289888882e-05,
      "loss": 1.3091,
      "mean_token_accuracy": 0.7556482389569282,
      "num_tokens": 1709245361.0,
      "step": 23780
    },
    {
      "entropy": 0.8908647835254669,
      "epoch": 6.484055600981193,
      "grad_norm": 0.12928390502929688,
      "learning_rate": 6.509668370439768e-05,
      "loss": 1.3178,
      "mean_token_accuracy": 0.7556312114000321,
      "num_tokens": 1709976801.0,
      "step": 23790
    },
    {
      "entropy": 0.8940265402197838,
      "epoch": 6.486781139275007,
      "grad_norm": 0.12529900670051575,
      "learning_rate": 6.506803905910603e-05,
      "loss": 1.3191,
      "mean_token_accuracy": 0.7518533378839493,
      "num_tokens": 1710686925.0,
      "step": 23800
    },
    {
      "entropy": 0.8857969135046005,
      "epoch": 6.48950667756882,
      "grad_norm": 0.1354072242975235,
      "learning_rate": 6.503938897335624e-05,
      "loss": 1.3156,
      "mean_token_accuracy": 0.755453372001648,
      "num_tokens": 1711400865.0,
      "step": 23810
    },
    {
      "entropy": 0.8843122676014901,
      "epoch": 6.492232215862633,
      "grad_norm": 0.12267336249351501,
      "learning_rate": 6.501073345749273e-05,
      "loss": 1.3061,
      "mean_token_accuracy": 0.7538071915507316,
      "num_tokens": 1712128288.0,
      "step": 23820
    },
    {
      "entropy": 0.8946838319301605,
      "epoch": 6.494957754156446,
      "grad_norm": 0.125118687748909,
      "learning_rate": 6.498207252186184e-05,
      "loss": 1.3199,
      "mean_token_accuracy": 0.7512657254934311,
      "num_tokens": 1712849010.0,
      "step": 23830
    },
    {
      "entropy": 0.8896793156862259,
      "epoch": 6.497683292450259,
      "grad_norm": 0.13046804070472717,
      "learning_rate": 6.495340617681188e-05,
      "loss": 1.3123,
      "mean_token_accuracy": 0.7554600208997726,
      "num_tokens": 1713573289.0,
      "step": 23840
    },
    {
      "entropy": 0.8903471857309342,
      "epoch": 6.500408830744072,
      "grad_norm": 0.14014749228954315,
      "learning_rate": 6.492473443269309e-05,
      "loss": 1.3121,
      "mean_token_accuracy": 0.7567742720246315,
      "num_tokens": 1714305119.0,
      "step": 23850
    },
    {
      "entropy": 0.8933516576886177,
      "epoch": 6.503134369037885,
      "grad_norm": 0.1295813024044037,
      "learning_rate": 6.48960572998577e-05,
      "loss": 1.3305,
      "mean_token_accuracy": 0.7485965028405189,
      "num_tokens": 1715010500.0,
      "step": 23860
    },
    {
      "entropy": 0.897659121453762,
      "epoch": 6.505859907331698,
      "grad_norm": 0.13633602857589722,
      "learning_rate": 6.486737478865982e-05,
      "loss": 1.3205,
      "mean_token_accuracy": 0.7503689274191856,
      "num_tokens": 1715724001.0,
      "step": 23870
    },
    {
      "entropy": 0.8984023123979569,
      "epoch": 6.508585445625511,
      "grad_norm": 0.13099007308483124,
      "learning_rate": 6.483868690945558e-05,
      "loss": 1.3235,
      "mean_token_accuracy": 0.7546491786837578,
      "num_tokens": 1716459393.0,
      "step": 23880
    },
    {
      "entropy": 0.8881630569696426,
      "epoch": 6.511310983919324,
      "grad_norm": 0.12748360633850098,
      "learning_rate": 6.480999367260298e-05,
      "loss": 1.3122,
      "mean_token_accuracy": 0.75551338493824,
      "num_tokens": 1717181110.0,
      "step": 23890
    },
    {
      "entropy": 0.8861339643597603,
      "epoch": 6.514036522213138,
      "grad_norm": 0.13338781893253326,
      "learning_rate": 6.4781295088462e-05,
      "loss": 1.3168,
      "mean_token_accuracy": 0.7524291262030601,
      "num_tokens": 1717882430.0,
      "step": 23900
    },
    {
      "entropy": 0.8883640363812446,
      "epoch": 6.5167620605069505,
      "grad_norm": 0.1292986422777176,
      "learning_rate": 6.475259116739454e-05,
      "loss": 1.3172,
      "mean_token_accuracy": 0.7540691047906876,
      "num_tokens": 1718593758.0,
      "step": 23910
    },
    {
      "entropy": 0.8899330779910087,
      "epoch": 6.5194875988007635,
      "grad_norm": 0.13770543038845062,
      "learning_rate": 6.472388191976441e-05,
      "loss": 1.317,
      "mean_token_accuracy": 0.7589975893497467,
      "num_tokens": 1719328185.0,
      "step": 23920
    },
    {
      "entropy": 0.8917457908391953,
      "epoch": 6.522213137094576,
      "grad_norm": 0.13328559696674347,
      "learning_rate": 6.469516735593737e-05,
      "loss": 1.318,
      "mean_token_accuracy": 0.7550568327307701,
      "num_tokens": 1720048314.0,
      "step": 23930
    },
    {
      "entropy": 0.8939030185341835,
      "epoch": 6.524938675388389,
      "grad_norm": 0.1289254128932953,
      "learning_rate": 6.466644748628107e-05,
      "loss": 1.3258,
      "mean_token_accuracy": 0.7536257997155189,
      "num_tokens": 1720756720.0,
      "step": 23940
    },
    {
      "entropy": 0.8894197911024093,
      "epoch": 6.527664213682202,
      "grad_norm": 0.1329088807106018,
      "learning_rate": 6.46377223211651e-05,
      "loss": 1.3161,
      "mean_token_accuracy": 0.7506936267018318,
      "num_tokens": 1721465576.0,
      "step": 23950
    },
    {
      "entropy": 0.8965966418385506,
      "epoch": 6.530389751976015,
      "grad_norm": 0.1288381814956665,
      "learning_rate": 6.460899187096095e-05,
      "loss": 1.3187,
      "mean_token_accuracy": 0.7482483148574829,
      "num_tokens": 1722179925.0,
      "step": 23960
    },
    {
      "entropy": 0.8917963176965713,
      "epoch": 6.533115290269828,
      "grad_norm": 0.12459111958742142,
      "learning_rate": 6.458025614604203e-05,
      "loss": 1.3103,
      "mean_token_accuracy": 0.7549076959490776,
      "num_tokens": 1722908032.0,
      "step": 23970
    },
    {
      "entropy": 0.8876068472862244,
      "epoch": 6.535840828563641,
      "grad_norm": 0.1401968002319336,
      "learning_rate": 6.455151515678363e-05,
      "loss": 1.3027,
      "mean_token_accuracy": 0.7560511589050293,
      "num_tokens": 1723631293.0,
      "step": 23980
    },
    {
      "entropy": 0.8865098312497139,
      "epoch": 6.538566366857454,
      "grad_norm": 0.12456918507814407,
      "learning_rate": 6.452276891356295e-05,
      "loss": 1.3125,
      "mean_token_accuracy": 0.7547439321875572,
      "num_tokens": 1724347625.0,
      "step": 23990
    },
    {
      "entropy": 0.8943420633673668,
      "epoch": 6.541291905151267,
      "grad_norm": 0.12663772702217102,
      "learning_rate": 6.449401742675913e-05,
      "loss": 1.3256,
      "mean_token_accuracy": 0.7505205154418946,
      "num_tokens": 1725053248.0,
      "step": 24000
    },
    {
      "entropy": 0.8970868140459061,
      "epoch": 6.54401744344508,
      "grad_norm": 0.12651976943016052,
      "learning_rate": 6.446526070675314e-05,
      "loss": 1.3213,
      "mean_token_accuracy": 0.7507851839065551,
      "num_tokens": 1725774801.0,
      "step": 24010
    },
    {
      "entropy": 0.8989195540547371,
      "epoch": 6.546742981738894,
      "grad_norm": 0.13023047149181366,
      "learning_rate": 6.443649876392785e-05,
      "loss": 1.3281,
      "mean_token_accuracy": 0.7508052751421929,
      "num_tokens": 1726487268.0,
      "step": 24020
    },
    {
      "entropy": 0.8936064541339874,
      "epoch": 6.549468520032707,
      "grad_norm": 0.13014456629753113,
      "learning_rate": 6.440773160866807e-05,
      "loss": 1.3149,
      "mean_token_accuracy": 0.7517767742276191,
      "num_tokens": 1727216659.0,
      "step": 24030
    },
    {
      "entropy": 0.8856271132826805,
      "epoch": 6.55219405832652,
      "grad_norm": 0.13365373015403748,
      "learning_rate": 6.437895925136043e-05,
      "loss": 1.3129,
      "mean_token_accuracy": 0.7560368105769157,
      "num_tokens": 1727943823.0,
      "step": 24040
    },
    {
      "entropy": 0.8843535616993904,
      "epoch": 6.554919596620333,
      "grad_norm": 0.12778621912002563,
      "learning_rate": 6.435018170239345e-05,
      "loss": 1.3056,
      "mean_token_accuracy": 0.75809406042099,
      "num_tokens": 1728669551.0,
      "step": 24050
    },
    {
      "entropy": 0.8948861703276634,
      "epoch": 6.557645134914146,
      "grad_norm": 0.130461186170578,
      "learning_rate": 6.432139897215757e-05,
      "loss": 1.317,
      "mean_token_accuracy": 0.7529885828495025,
      "num_tokens": 1729389158.0,
      "step": 24060
    },
    {
      "entropy": 0.8886276677250862,
      "epoch": 6.5603706732079585,
      "grad_norm": 0.13357917964458466,
      "learning_rate": 6.429261107104506e-05,
      "loss": 1.3154,
      "mean_token_accuracy": 0.7550659969449043,
      "num_tokens": 1730105326.0,
      "step": 24070
    },
    {
      "entropy": 0.8856066077947616,
      "epoch": 6.5630962115017715,
      "grad_norm": 0.12202835828065872,
      "learning_rate": 6.426381800945003e-05,
      "loss": 1.322,
      "mean_token_accuracy": 0.7514744594693183,
      "num_tokens": 1730810191.0,
      "step": 24080
    },
    {
      "entropy": 0.8879060074687004,
      "epoch": 6.565821749795584,
      "grad_norm": 0.1253044158220291,
      "learning_rate": 6.423501979776851e-05,
      "loss": 1.3161,
      "mean_token_accuracy": 0.7507121905684471,
      "num_tokens": 1731515682.0,
      "step": 24090
    },
    {
      "entropy": 0.8891847297549248,
      "epoch": 6.568547288089397,
      "grad_norm": 0.13624908030033112,
      "learning_rate": 6.420621644639838e-05,
      "loss": 1.3143,
      "mean_token_accuracy": 0.7506772324442863,
      "num_tokens": 1732224391.0,
      "step": 24100
    },
    {
      "entropy": 0.8838733717799186,
      "epoch": 6.57127282638321,
      "grad_norm": 0.13081113994121552,
      "learning_rate": 6.417740796573935e-05,
      "loss": 1.3093,
      "mean_token_accuracy": 0.7561607211828232,
      "num_tokens": 1732948736.0,
      "step": 24110
    },
    {
      "entropy": 0.8956152871251106,
      "epoch": 6.573998364677024,
      "grad_norm": 0.11970707774162292,
      "learning_rate": 6.414859436619299e-05,
      "loss": 1.3265,
      "mean_token_accuracy": 0.7495107665657997,
      "num_tokens": 1733662862.0,
      "step": 24120
    },
    {
      "entropy": 0.8898704186081886,
      "epoch": 6.576723902970837,
      "grad_norm": 0.22617320716381073,
      "learning_rate": 6.411977565816274e-05,
      "loss": 1.3127,
      "mean_token_accuracy": 0.7540691748261452,
      "num_tokens": 1734386723.0,
      "step": 24130
    },
    {
      "entropy": 0.8996606901288032,
      "epoch": 6.57944944126465,
      "grad_norm": 0.13188959658145905,
      "learning_rate": 6.409095185205385e-05,
      "loss": 1.3279,
      "mean_token_accuracy": 0.7504490211606025,
      "num_tokens": 1735107352.0,
      "step": 24140
    },
    {
      "entropy": 0.8828933015465736,
      "epoch": 6.582174979558463,
      "grad_norm": 0.12779057025909424,
      "learning_rate": 6.406212295827342e-05,
      "loss": 1.3121,
      "mean_token_accuracy": 0.7595152884721756,
      "num_tokens": 1735836339.0,
      "step": 24150
    },
    {
      "entropy": 0.8865995883941651,
      "epoch": 6.584900517852276,
      "grad_norm": 0.13406582176685333,
      "learning_rate": 6.403328898723043e-05,
      "loss": 1.3055,
      "mean_token_accuracy": 0.7550941169261932,
      "num_tokens": 1736557165.0,
      "step": 24160
    },
    {
      "entropy": 0.8959143862128258,
      "epoch": 6.587626056146089,
      "grad_norm": 0.12159795314073563,
      "learning_rate": 6.400444994933562e-05,
      "loss": 1.3164,
      "mean_token_accuracy": 0.7527622133493423,
      "num_tokens": 1737283974.0,
      "step": 24170
    },
    {
      "entropy": 0.8952702283859253,
      "epoch": 6.590351594439902,
      "grad_norm": 0.1343999058008194,
      "learning_rate": 6.397560585500162e-05,
      "loss": 1.3241,
      "mean_token_accuracy": 0.7504002094268799,
      "num_tokens": 1737995468.0,
      "step": 24180
    },
    {
      "entropy": 0.880812081694603,
      "epoch": 6.593077132733715,
      "grad_norm": 0.12571381032466888,
      "learning_rate": 6.394675671464285e-05,
      "loss": 1.3063,
      "mean_token_accuracy": 0.7595406115055084,
      "num_tokens": 1738737475.0,
      "step": 24190
    },
    {
      "entropy": 0.8856984466314316,
      "epoch": 6.595802671027528,
      "grad_norm": 0.12951508164405823,
      "learning_rate": 6.391790253867558e-05,
      "loss": 1.3019,
      "mean_token_accuracy": 0.7567184373736382,
      "num_tokens": 1739467299.0,
      "step": 24200
    },
    {
      "entropy": 0.8935243427753449,
      "epoch": 6.598528209321341,
      "grad_norm": 0.11822971701622009,
      "learning_rate": 6.388904333751787e-05,
      "loss": 1.3149,
      "mean_token_accuracy": 0.7509625971317291,
      "num_tokens": 1740182846.0,
      "step": 24210
    },
    {
      "entropy": 0.885768061876297,
      "epoch": 6.601253747615154,
      "grad_norm": 0.12649142742156982,
      "learning_rate": 6.386017912158962e-05,
      "loss": 1.3177,
      "mean_token_accuracy": 0.756460054218769,
      "num_tokens": 1740913451.0,
      "step": 24220
    },
    {
      "entropy": 0.8956640034914016,
      "epoch": 6.6039792859089665,
      "grad_norm": 0.13004878163337708,
      "learning_rate": 6.383130990131252e-05,
      "loss": 1.3257,
      "mean_token_accuracy": 0.7497450560331345,
      "num_tokens": 1741622837.0,
      "step": 24230
    },
    {
      "entropy": 0.8826272994279861,
      "epoch": 6.60670482420278,
      "grad_norm": 0.12117267400026321,
      "learning_rate": 6.380243568711006e-05,
      "loss": 1.3083,
      "mean_token_accuracy": 0.7582021147012711,
      "num_tokens": 1742351492.0,
      "step": 24240
    },
    {
      "entropy": 0.8956345677375793,
      "epoch": 6.609430362496593,
      "grad_norm": 0.1367512345314026,
      "learning_rate": 6.377355648940759e-05,
      "loss": 1.3223,
      "mean_token_accuracy": 0.7494165197014808,
      "num_tokens": 1743053734.0,
      "step": 24250
    },
    {
      "entropy": 0.8893290594220161,
      "epoch": 6.612155900790406,
      "grad_norm": 0.12484463304281235,
      "learning_rate": 6.37446723186322e-05,
      "loss": 1.3049,
      "mean_token_accuracy": 0.7561960011720658,
      "num_tokens": 1743783093.0,
      "step": 24260
    },
    {
      "entropy": 0.893658597767353,
      "epoch": 6.614881439084219,
      "grad_norm": 0.12526802718639374,
      "learning_rate": 6.371578318521275e-05,
      "loss": 1.312,
      "mean_token_accuracy": 0.7512623444199562,
      "num_tokens": 1744503404.0,
      "step": 24270
    },
    {
      "entropy": 0.8851838201284409,
      "epoch": 6.617606977378032,
      "grad_norm": 0.14824999868869781,
      "learning_rate": 6.368688909958001e-05,
      "loss": 1.3101,
      "mean_token_accuracy": 0.7554869264364242,
      "num_tokens": 1745213421.0,
      "step": 24280
    },
    {
      "entropy": 0.8950522854924202,
      "epoch": 6.620332515671845,
      "grad_norm": 0.13106846809387207,
      "learning_rate": 6.36579900721664e-05,
      "loss": 1.327,
      "mean_token_accuracy": 0.7499068573117256,
      "num_tokens": 1745925812.0,
      "step": 24290
    },
    {
      "entropy": 0.8931015208363533,
      "epoch": 6.623058053965658,
      "grad_norm": 0.12991878390312195,
      "learning_rate": 6.362908611340621e-05,
      "loss": 1.3193,
      "mean_token_accuracy": 0.752657313644886,
      "num_tokens": 1746653727.0,
      "step": 24300
    },
    {
      "entropy": 0.8938193216919899,
      "epoch": 6.625783592259471,
      "grad_norm": 0.12958885729312897,
      "learning_rate": 6.36001772337355e-05,
      "loss": 1.32,
      "mean_token_accuracy": 0.7505113437771798,
      "num_tokens": 1747360152.0,
      "step": 24310
    },
    {
      "entropy": 0.8950100347399712,
      "epoch": 6.628509130553284,
      "grad_norm": 0.11723659932613373,
      "learning_rate": 6.357126344359209e-05,
      "loss": 1.3208,
      "mean_token_accuracy": 0.7511440187692642,
      "num_tokens": 1748084047.0,
      "step": 24320
    },
    {
      "entropy": 0.8897304311394691,
      "epoch": 6.631234668847098,
      "grad_norm": 0.1372891217470169,
      "learning_rate": 6.354234475341556e-05,
      "loss": 1.3188,
      "mean_token_accuracy": 0.7515660494565963,
      "num_tokens": 1748790112.0,
      "step": 24330
    },
    {
      "entropy": 0.8884767234325409,
      "epoch": 6.633960207140911,
      "grad_norm": 0.12608219683170319,
      "learning_rate": 6.35134211736473e-05,
      "loss": 1.3138,
      "mean_token_accuracy": 0.7541348546743393,
      "num_tokens": 1749507751.0,
      "step": 24340
    },
    {
      "entropy": 0.8946855247020722,
      "epoch": 6.636685745434724,
      "grad_norm": 0.12446733564138412,
      "learning_rate": 6.348449271473042e-05,
      "loss": 1.3181,
      "mean_token_accuracy": 0.7499898329377175,
      "num_tokens": 1750216543.0,
      "step": 24350
    },
    {
      "entropy": 0.8841793164610863,
      "epoch": 6.639411283728537,
      "grad_norm": 0.13644251227378845,
      "learning_rate": 6.345555938710982e-05,
      "loss": 1.3087,
      "mean_token_accuracy": 0.7546942919492722,
      "num_tokens": 1750928991.0,
      "step": 24360
    },
    {
      "entropy": 0.8819687396287919,
      "epoch": 6.64213682202235,
      "grad_norm": 0.13751031458377838,
      "learning_rate": 6.342662120123218e-05,
      "loss": 1.318,
      "mean_token_accuracy": 0.7530433446168899,
      "num_tokens": 1751630163.0,
      "step": 24370
    },
    {
      "entropy": 0.8908970817923546,
      "epoch": 6.6448623603161625,
      "grad_norm": 0.12846535444259644,
      "learning_rate": 6.339767816754585e-05,
      "loss": 1.317,
      "mean_token_accuracy": 0.7512157216668129,
      "num_tokens": 1752349451.0,
      "step": 24380
    },
    {
      "entropy": 0.895119409263134,
      "epoch": 6.6475878986099755,
      "grad_norm": 0.13087886571884155,
      "learning_rate": 6.336873029650104e-05,
      "loss": 1.3154,
      "mean_token_accuracy": 0.7545461416244507,
      "num_tokens": 1753086086.0,
      "step": 24390
    },
    {
      "entropy": 0.8943515375256539,
      "epoch": 6.650313436903788,
      "grad_norm": 0.1285870373249054,
      "learning_rate": 6.333977759854962e-05,
      "loss": 1.3122,
      "mean_token_accuracy": 0.7522528052330018,
      "num_tokens": 1753811823.0,
      "step": 24400
    },
    {
      "entropy": 0.8933129757642746,
      "epoch": 6.653038975197601,
      "grad_norm": 0.1275702714920044,
      "learning_rate": 6.331082008414525e-05,
      "loss": 1.3158,
      "mean_token_accuracy": 0.7519129410386085,
      "num_tokens": 1754536761.0,
      "step": 24410
    },
    {
      "entropy": 0.8880542337894439,
      "epoch": 6.655764513491414,
      "grad_norm": 0.12732753157615662,
      "learning_rate": 6.328185776374331e-05,
      "loss": 1.3131,
      "mean_token_accuracy": 0.7550013184547424,
      "num_tokens": 1755250630.0,
      "step": 24420
    },
    {
      "entropy": 0.8821530431509018,
      "epoch": 6.658490051785227,
      "grad_norm": 0.12260854989290237,
      "learning_rate": 6.325289064780094e-05,
      "loss": 1.3143,
      "mean_token_accuracy": 0.752858467400074,
      "num_tokens": 1755956287.0,
      "step": 24430
    },
    {
      "entropy": 0.8928753837943078,
      "epoch": 6.66121559007904,
      "grad_norm": 0.27121585607528687,
      "learning_rate": 6.322391874677695e-05,
      "loss": 1.3188,
      "mean_token_accuracy": 0.7530712425708771,
      "num_tokens": 1756682135.0,
      "step": 24440
    },
    {
      "entropy": 0.8917052865028381,
      "epoch": 6.663941128372854,
      "grad_norm": 0.12325097620487213,
      "learning_rate": 6.319494207113195e-05,
      "loss": 1.3142,
      "mean_token_accuracy": 0.7508950963616371,
      "num_tokens": 1757395883.0,
      "step": 24450
    },
    {
      "entropy": 0.8906190350651741,
      "epoch": 6.666666666666667,
      "grad_norm": 0.12949134409427643,
      "learning_rate": 6.316596063132822e-05,
      "loss": 1.319,
      "mean_token_accuracy": 0.7552656531333923,
      "num_tokens": 1758120405.0,
      "step": 24460
    },
    {
      "entropy": 0.8926933661103249,
      "epoch": 6.66939220496048,
      "grad_norm": 0.12961310148239136,
      "learning_rate": 6.313697443782982e-05,
      "loss": 1.322,
      "mean_token_accuracy": 0.7506764069199562,
      "num_tokens": 1758828318.0,
      "step": 24470
    },
    {
      "entropy": 0.8932582020759583,
      "epoch": 6.672117743254293,
      "grad_norm": 0.1281840056180954,
      "learning_rate": 6.310798350110246e-05,
      "loss": 1.319,
      "mean_token_accuracy": 0.7502627789974212,
      "num_tokens": 1759532559.0,
      "step": 24480
    },
    {
      "entropy": 0.8890052795410156,
      "epoch": 6.674843281548106,
      "grad_norm": 0.1319226175546646,
      "learning_rate": 6.30789878316136e-05,
      "loss": 1.3175,
      "mean_token_accuracy": 0.7541633531451225,
      "num_tokens": 1760243785.0,
      "step": 24490
    },
    {
      "entropy": 0.8888242825865745,
      "epoch": 6.677568819841919,
      "grad_norm": 0.12243334203958511,
      "learning_rate": 6.304998743983239e-05,
      "loss": 1.3163,
      "mean_token_accuracy": 0.7570770010352135,
      "num_tokens": 1760972586.0,
      "step": 24500
    },
    {
      "entropy": 0.8893096715211868,
      "epoch": 6.680294358135732,
      "grad_norm": 0.13131950795650482,
      "learning_rate": 6.302098233622972e-05,
      "loss": 1.3157,
      "mean_token_accuracy": 0.7527543500065803,
      "num_tokens": 1761689047.0,
      "step": 24510
    },
    {
      "entropy": 0.887695050239563,
      "epoch": 6.683019896429545,
      "grad_norm": 0.1326379030942917,
      "learning_rate": 6.299197253127815e-05,
      "loss": 1.3146,
      "mean_token_accuracy": 0.7535473465919494,
      "num_tokens": 1762411702.0,
      "step": 24520
    },
    {
      "entropy": 0.8963560268282891,
      "epoch": 6.685745434723358,
      "grad_norm": 0.13111519813537598,
      "learning_rate": 6.296295803545191e-05,
      "loss": 1.3281,
      "mean_token_accuracy": 0.7485727503895759,
      "num_tokens": 1763116185.0,
      "step": 24530
    },
    {
      "entropy": 0.8973544210195541,
      "epoch": 6.6884709730171705,
      "grad_norm": 0.13762107491493225,
      "learning_rate": 6.293393885922702e-05,
      "loss": 1.3202,
      "mean_token_accuracy": 0.7498584285378456,
      "num_tokens": 1763833304.0,
      "step": 24540
    },
    {
      "entropy": 0.8898538023233413,
      "epoch": 6.691196511310984,
      "grad_norm": 0.13277111947536469,
      "learning_rate": 6.290491501308108e-05,
      "loss": 1.3106,
      "mean_token_accuracy": 0.7581757709383965,
      "num_tokens": 1764568051.0,
      "step": 24550
    },
    {
      "entropy": 0.8891453593969345,
      "epoch": 6.693922049604797,
      "grad_norm": 0.12729622423648834,
      "learning_rate": 6.287588650749343e-05,
      "loss": 1.3131,
      "mean_token_accuracy": 0.7540027663111687,
      "num_tokens": 1765300001.0,
      "step": 24560
    },
    {
      "entropy": 0.895088429749012,
      "epoch": 6.69664758789861,
      "grad_norm": 0.13212324678897858,
      "learning_rate": 6.284685335294511e-05,
      "loss": 1.3239,
      "mean_token_accuracy": 0.7556274324655533,
      "num_tokens": 1766016154.0,
      "step": 24570
    },
    {
      "entropy": 0.8938517928123474,
      "epoch": 6.699373126192423,
      "grad_norm": 0.14164453744888306,
      "learning_rate": 6.281781555991878e-05,
      "loss": 1.3237,
      "mean_token_accuracy": 0.7555088177323341,
      "num_tokens": 1766743771.0,
      "step": 24580
    },
    {
      "entropy": 0.8876852571964264,
      "epoch": 6.702098664486236,
      "grad_norm": 0.12492102384567261,
      "learning_rate": 6.278877313889884e-05,
      "loss": 1.3143,
      "mean_token_accuracy": 0.7587466761469841,
      "num_tokens": 1767475491.0,
      "step": 24590
    },
    {
      "entropy": 0.8952235966920853,
      "epoch": 6.704824202780049,
      "grad_norm": 0.12633070349693298,
      "learning_rate": 6.275972610037136e-05,
      "loss": 1.3288,
      "mean_token_accuracy": 0.7498063012957573,
      "num_tokens": 1768176601.0,
      "step": 24600
    },
    {
      "entropy": 0.8925411343574524,
      "epoch": 6.707549741073862,
      "grad_norm": 0.13363733887672424,
      "learning_rate": 6.273067445482396e-05,
      "loss": 1.3299,
      "mean_token_accuracy": 0.7524307861924171,
      "num_tokens": 1768881833.0,
      "step": 24610
    },
    {
      "entropy": 0.8888502702116966,
      "epoch": 6.710275279367675,
      "grad_norm": 0.13241320848464966,
      "learning_rate": 6.270161821274607e-05,
      "loss": 1.3151,
      "mean_token_accuracy": 0.7526521891355514,
      "num_tokens": 1769597883.0,
      "step": 24620
    },
    {
      "entropy": 0.8879201605916023,
      "epoch": 6.713000817661488,
      "grad_norm": 0.12263783067464828,
      "learning_rate": 6.267255738462874e-05,
      "loss": 1.3126,
      "mean_token_accuracy": 0.7539681926369667,
      "num_tokens": 1770310587.0,
      "step": 24630
    },
    {
      "entropy": 0.8861521303653717,
      "epoch": 6.715726355955301,
      "grad_norm": 0.1356944441795349,
      "learning_rate": 6.26434919809646e-05,
      "loss": 1.3173,
      "mean_token_accuracy": 0.7576313450932503,
      "num_tokens": 1771030794.0,
      "step": 24640
    },
    {
      "entropy": 0.8939735934138298,
      "epoch": 6.718451894249114,
      "grad_norm": 0.128499835729599,
      "learning_rate": 6.261442201224803e-05,
      "loss": 1.3242,
      "mean_token_accuracy": 0.756110742688179,
      "num_tokens": 1771750478.0,
      "step": 24650
    },
    {
      "entropy": 0.8817349553108216,
      "epoch": 6.721177432542927,
      "grad_norm": 0.13707764446735382,
      "learning_rate": 6.2585347488975e-05,
      "loss": 1.3123,
      "mean_token_accuracy": 0.7497487619519234,
      "num_tokens": 1772452228.0,
      "step": 24660
    },
    {
      "entropy": 0.8883233219385147,
      "epoch": 6.723902970836741,
      "grad_norm": 0.1376660317182541,
      "learning_rate": 6.255626842164318e-05,
      "loss": 1.3183,
      "mean_token_accuracy": 0.754699669778347,
      "num_tokens": 1773165534.0,
      "step": 24670
    },
    {
      "entropy": 0.8869135260581971,
      "epoch": 6.7266285091305535,
      "grad_norm": 0.1446562558412552,
      "learning_rate": 6.252718482075177e-05,
      "loss": 1.3174,
      "mean_token_accuracy": 0.7535565987229347,
      "num_tokens": 1773874549.0,
      "step": 24680
    },
    {
      "entropy": 0.9035732954740524,
      "epoch": 6.7293540474243665,
      "grad_norm": 0.12701532244682312,
      "learning_rate": 6.249809669680174e-05,
      "loss": 1.3247,
      "mean_token_accuracy": 0.7481285810470581,
      "num_tokens": 1774594622.0,
      "step": 24690
    },
    {
      "entropy": 0.8916135907173157,
      "epoch": 6.732079585718179,
      "grad_norm": 0.12978671491146088,
      "learning_rate": 6.246900406029558e-05,
      "loss": 1.3152,
      "mean_token_accuracy": 0.7495876282453537,
      "num_tokens": 1775296615.0,
      "step": 24700
    },
    {
      "entropy": 0.8902298852801322,
      "epoch": 6.734805124011992,
      "grad_norm": 0.13260269165039062,
      "learning_rate": 6.243990692173753e-05,
      "loss": 1.3091,
      "mean_token_accuracy": 0.7554959207773209,
      "num_tokens": 1776023118.0,
      "step": 24710
    },
    {
      "entropy": 0.883197845518589,
      "epoch": 6.737530662305805,
      "grad_norm": 0.12556543946266174,
      "learning_rate": 6.241080529163335e-05,
      "loss": 1.3082,
      "mean_token_accuracy": 0.7563388183712959,
      "num_tokens": 1776750009.0,
      "step": 24720
    },
    {
      "entropy": 0.8866877958178521,
      "epoch": 6.740256200599618,
      "grad_norm": 0.13022150099277496,
      "learning_rate": 6.238169918049045e-05,
      "loss": 1.3174,
      "mean_token_accuracy": 0.7532852962613106,
      "num_tokens": 1777471682.0,
      "step": 24730
    },
    {
      "entropy": 0.8915566191077232,
      "epoch": 6.742981738893431,
      "grad_norm": 0.13103778660297394,
      "learning_rate": 6.235258859881787e-05,
      "loss": 1.3129,
      "mean_token_accuracy": 0.7541903033852577,
      "num_tokens": 1778197294.0,
      "step": 24740
    },
    {
      "entropy": 0.8954208865761757,
      "epoch": 6.745707277187244,
      "grad_norm": 0.12838414311408997,
      "learning_rate": 6.232347355712629e-05,
      "loss": 1.3147,
      "mean_token_accuracy": 0.7506286948919296,
      "num_tokens": 1778909783.0,
      "step": 24750
    },
    {
      "entropy": 0.8895563051104546,
      "epoch": 6.748432815481058,
      "grad_norm": 0.12716424465179443,
      "learning_rate": 6.229435406592794e-05,
      "loss": 1.3104,
      "mean_token_accuracy": 0.7477434396743774,
      "num_tokens": 1779613678.0,
      "step": 24760
    },
    {
      "entropy": 0.8930724635720253,
      "epoch": 6.751158353774871,
      "grad_norm": 0.14486993849277496,
      "learning_rate": 6.226523013573668e-05,
      "loss": 1.3197,
      "mean_token_accuracy": 0.7560727789998054,
      "num_tokens": 1780346417.0,
      "step": 24770
    },
    {
      "entropy": 0.8904518365859986,
      "epoch": 6.753883892068684,
      "grad_norm": 0.12527410686016083,
      "learning_rate": 6.223610177706803e-05,
      "loss": 1.3226,
      "mean_token_accuracy": 0.7507403671741486,
      "num_tokens": 1781052552.0,
      "step": 24780
    },
    {
      "entropy": 0.8915664985775947,
      "epoch": 6.756609430362497,
      "grad_norm": 0.14312466979026794,
      "learning_rate": 6.2206969000439e-05,
      "loss": 1.3168,
      "mean_token_accuracy": 0.7518466025590896,
      "num_tokens": 1781770239.0,
      "step": 24790
    },
    {
      "entropy": 0.8863757699728012,
      "epoch": 6.75933496865631,
      "grad_norm": 0.13256804645061493,
      "learning_rate": 6.21778318163683e-05,
      "loss": 1.3163,
      "mean_token_accuracy": 0.7572139993309974,
      "num_tokens": 1782493481.0,
      "step": 24800
    },
    {
      "entropy": 0.8955301865935326,
      "epoch": 6.762060506950123,
      "grad_norm": 0.12993404269218445,
      "learning_rate": 6.214869023537617e-05,
      "loss": 1.3216,
      "mean_token_accuracy": 0.7534404128789902,
      "num_tokens": 1783210604.0,
      "step": 24810
    },
    {
      "entropy": 0.8905673056840897,
      "epoch": 6.764786045243936,
      "grad_norm": 0.13321179151535034,
      "learning_rate": 6.211954426798445e-05,
      "loss": 1.3189,
      "mean_token_accuracy": 0.7539059400558472,
      "num_tokens": 1783937145.0,
      "step": 24820
    },
    {
      "entropy": 0.888870345056057,
      "epoch": 6.767511583537749,
      "grad_norm": 0.12881068885326385,
      "learning_rate": 6.209039392471655e-05,
      "loss": 1.3109,
      "mean_token_accuracy": 0.7551013827323914,
      "num_tokens": 1784654847.0,
      "step": 24830
    },
    {
      "entropy": 0.8842212155461311,
      "epoch": 6.770237121831562,
      "grad_norm": 0.1341342031955719,
      "learning_rate": 6.206123921609752e-05,
      "loss": 1.3169,
      "mean_token_accuracy": 0.7548576384782791,
      "num_tokens": 1785368803.0,
      "step": 24840
    },
    {
      "entropy": 0.8953014209866523,
      "epoch": 6.7729626601253745,
      "grad_norm": 0.12916676700115204,
      "learning_rate": 6.20320801526539e-05,
      "loss": 1.3152,
      "mean_token_accuracy": 0.7519174054265022,
      "num_tokens": 1786087885.0,
      "step": 24850
    },
    {
      "entropy": 0.8924761295318604,
      "epoch": 6.7756881984191875,
      "grad_norm": 0.1352209597826004,
      "learning_rate": 6.200291674491389e-05,
      "loss": 1.3171,
      "mean_token_accuracy": 0.7516786128282547,
      "num_tokens": 1786794685.0,
      "step": 24860
    },
    {
      "entropy": 0.8978717163205147,
      "epoch": 6.778413736713,
      "grad_norm": 0.13338683545589447,
      "learning_rate": 6.197374900340717e-05,
      "loss": 1.3268,
      "mean_token_accuracy": 0.7472734853625298,
      "num_tokens": 1787495432.0,
      "step": 24870
    },
    {
      "entropy": 0.8823297992348671,
      "epoch": 6.781139275006813,
      "grad_norm": 0.14823059737682343,
      "learning_rate": 6.194457693866507e-05,
      "loss": 1.3046,
      "mean_token_accuracy": 0.760610094666481,
      "num_tokens": 1788234159.0,
      "step": 24880
    },
    {
      "entropy": 0.8852872222661972,
      "epoch": 6.783864813300627,
      "grad_norm": 0.12477268278598785,
      "learning_rate": 6.191540056122039e-05,
      "loss": 1.3108,
      "mean_token_accuracy": 0.7568852975964546,
      "num_tokens": 1788952975.0,
      "step": 24890
    },
    {
      "entropy": 0.889084380865097,
      "epoch": 6.78659035159444,
      "grad_norm": 0.1310306191444397,
      "learning_rate": 6.188621988160759e-05,
      "loss": 1.3193,
      "mean_token_accuracy": 0.7547228544950485,
      "num_tokens": 1789667292.0,
      "step": 24900
    },
    {
      "entropy": 0.8856345131993294,
      "epoch": 6.789315889888253,
      "grad_norm": 0.13499148190021515,
      "learning_rate": 6.185703491036259e-05,
      "loss": 1.31,
      "mean_token_accuracy": 0.7545298427343369,
      "num_tokens": 1790393082.0,
      "step": 24910
    },
    {
      "entropy": 0.8889864951372146,
      "epoch": 6.792041428182066,
      "grad_norm": 0.13105006515979767,
      "learning_rate": 6.182784565802294e-05,
      "loss": 1.3076,
      "mean_token_accuracy": 0.7521557480096817,
      "num_tokens": 1791098703.0,
      "step": 24920
    },
    {
      "entropy": 0.8832415670156479,
      "epoch": 6.794766966475879,
      "grad_norm": 0.13367614150047302,
      "learning_rate": 6.179865213512766e-05,
      "loss": 1.3221,
      "mean_token_accuracy": 0.7522967308759689,
      "num_tokens": 1791802322.0,
      "step": 24930
    },
    {
      "entropy": 0.8917075216770172,
      "epoch": 6.797492504769692,
      "grad_norm": 0.12641644477844238,
      "learning_rate": 6.176945435221736e-05,
      "loss": 1.3228,
      "mean_token_accuracy": 0.7473835825920105,
      "num_tokens": 1792498190.0,
      "step": 24940
    },
    {
      "entropy": 0.8921925216913223,
      "epoch": 6.800218043063505,
      "grad_norm": 0.1273316591978073,
      "learning_rate": 6.174025231983416e-05,
      "loss": 1.3155,
      "mean_token_accuracy": 0.7546837568283081,
      "num_tokens": 1793218782.0,
      "step": 24950
    },
    {
      "entropy": 0.8941346377134323,
      "epoch": 6.802943581357318,
      "grad_norm": 0.1447201818227768,
      "learning_rate": 6.171104604852179e-05,
      "loss": 1.3222,
      "mean_token_accuracy": 0.7505118101835251,
      "num_tokens": 1793928192.0,
      "step": 24960
    },
    {
      "entropy": 0.8849840000271797,
      "epoch": 6.805669119651131,
      "grad_norm": 0.13333961367607117,
      "learning_rate": 6.168183554882538e-05,
      "loss": 1.3141,
      "mean_token_accuracy": 0.7556151181459427,
      "num_tokens": 1794652393.0,
      "step": 24970
    },
    {
      "entropy": 0.8927885562181472,
      "epoch": 6.808394657944945,
      "grad_norm": 0.13098543882369995,
      "learning_rate": 6.165262083129169e-05,
      "loss": 1.3183,
      "mean_token_accuracy": 0.7493065282702446,
      "num_tokens": 1795376456.0,
      "step": 24980
    },
    {
      "entropy": 0.8905992582440376,
      "epoch": 6.8111201962387575,
      "grad_norm": 0.13117153942584991,
      "learning_rate": 6.162340190646896e-05,
      "loss": 1.3244,
      "mean_token_accuracy": 0.7521332427859306,
      "num_tokens": 1796081411.0,
      "step": 24990
    },
    {
      "entropy": 0.890515461564064,
      "epoch": 6.8138457345325705,
      "grad_norm": 0.1338358074426651,
      "learning_rate": 6.159417878490697e-05,
      "loss": 1.3226,
      "mean_token_accuracy": 0.7561894357204437,
      "num_tokens": 1796793379.0,
      "step": 25000
    },
    {
      "entropy": 0.8920251533389092,
      "epoch": 6.816571272826383,
      "grad_norm": 0.1321912556886673,
      "learning_rate": 6.1564951477157e-05,
      "loss": 1.317,
      "mean_token_accuracy": 0.7513139307498932,
      "num_tokens": 1797508363.0,
      "step": 25010
    },
    {
      "entropy": 0.8878983482718468,
      "epoch": 6.819296811120196,
      "grad_norm": 0.12442489713430405,
      "learning_rate": 6.153571999377185e-05,
      "loss": 1.3196,
      "mean_token_accuracy": 0.7533785417675972,
      "num_tokens": 1798225244.0,
      "step": 25020
    },
    {
      "entropy": 0.8908624187111854,
      "epoch": 6.822022349414009,
      "grad_norm": 0.12250331789255142,
      "learning_rate": 6.150648434530581e-05,
      "loss": 1.3141,
      "mean_token_accuracy": 0.7504842028021812,
      "num_tokens": 1798932347.0,
      "step": 25030
    },
    {
      "entropy": 0.8861579939723014,
      "epoch": 6.824747887707822,
      "grad_norm": 0.12292330712080002,
      "learning_rate": 6.147724454231471e-05,
      "loss": 1.3195,
      "mean_token_accuracy": 0.7477668762207031,
      "num_tokens": 1799632761.0,
      "step": 25040
    },
    {
      "entropy": 0.8902771726250649,
      "epoch": 6.827473426001635,
      "grad_norm": 0.12106510996818542,
      "learning_rate": 6.144800059535584e-05,
      "loss": 1.3157,
      "mean_token_accuracy": 0.7549340993165969,
      "num_tokens": 1800364884.0,
      "step": 25050
    },
    {
      "entropy": 0.8949364453554154,
      "epoch": 6.830198964295448,
      "grad_norm": 0.12179964780807495,
      "learning_rate": 6.1418752514988e-05,
      "loss": 1.3219,
      "mean_token_accuracy": 0.7526259392499923,
      "num_tokens": 1801082374.0,
      "step": 25060
    },
    {
      "entropy": 0.8969263181090354,
      "epoch": 6.832924502589261,
      "grad_norm": 0.1304628849029541,
      "learning_rate": 6.13895003117715e-05,
      "loss": 1.3234,
      "mean_token_accuracy": 0.7511690348386765,
      "num_tokens": 1801800134.0,
      "step": 25070
    },
    {
      "entropy": 0.8968133956193924,
      "epoch": 6.835650040883074,
      "grad_norm": 0.12084237486124039,
      "learning_rate": 6.13602439962681e-05,
      "loss": 1.3172,
      "mean_token_accuracy": 0.7497126936912537,
      "num_tokens": 1802521122.0,
      "step": 25080
    },
    {
      "entropy": 0.8833630070090294,
      "epoch": 6.838375579176887,
      "grad_norm": 0.1426776647567749,
      "learning_rate": 6.133098357904111e-05,
      "loss": 1.314,
      "mean_token_accuracy": 0.7548305094242096,
      "num_tokens": 1803233108.0,
      "step": 25090
    },
    {
      "entropy": 0.8972644343972206,
      "epoch": 6.841101117470701,
      "grad_norm": 0.12756995856761932,
      "learning_rate": 6.130171907065528e-05,
      "loss": 1.3207,
      "mean_token_accuracy": 0.7509740442037582,
      "num_tokens": 1803943588.0,
      "step": 25100
    },
    {
      "entropy": 0.8945842295885086,
      "epoch": 6.843826655764514,
      "grad_norm": 0.12099883705377579,
      "learning_rate": 6.127245048167677e-05,
      "loss": 1.3134,
      "mean_token_accuracy": 0.7521846234798432,
      "num_tokens": 1804663812.0,
      "step": 25110
    },
    {
      "entropy": 0.8933159589767456,
      "epoch": 6.846552194058327,
      "grad_norm": 0.13169211149215698,
      "learning_rate": 6.124317782267334e-05,
      "loss": 1.323,
      "mean_token_accuracy": 0.7487967729568481,
      "num_tokens": 1805361304.0,
      "step": 25120
    },
    {
      "entropy": 0.8939870715141296,
      "epoch": 6.84927773235214,
      "grad_norm": 0.13507576286792755,
      "learning_rate": 6.121390110421417e-05,
      "loss": 1.3242,
      "mean_token_accuracy": 0.7542021676898003,
      "num_tokens": 1806090119.0,
      "step": 25130
    },
    {
      "entropy": 0.8863414153456688,
      "epoch": 6.852003270645953,
      "grad_norm": 0.14156360924243927,
      "learning_rate": 6.118462033686986e-05,
      "loss": 1.3115,
      "mean_token_accuracy": 0.7529726713895798,
      "num_tokens": 1806806945.0,
      "step": 25140
    },
    {
      "entropy": 0.8881063982844353,
      "epoch": 6.8547288089397655,
      "grad_norm": 0.12910422682762146,
      "learning_rate": 6.115533553121253e-05,
      "loss": 1.3182,
      "mean_token_accuracy": 0.7578499868512154,
      "num_tokens": 1807530800.0,
      "step": 25150
    },
    {
      "entropy": 0.8826519504189492,
      "epoch": 6.8574543472335785,
      "grad_norm": 0.12877538800239563,
      "learning_rate": 6.112604669781572e-05,
      "loss": 1.307,
      "mean_token_accuracy": 0.7581726223230362,
      "num_tokens": 1808251647.0,
      "step": 25160
    },
    {
      "entropy": 0.879099203646183,
      "epoch": 6.860179885527391,
      "grad_norm": 0.12529128789901733,
      "learning_rate": 6.109675384725446e-05,
      "loss": 1.3147,
      "mean_token_accuracy": 0.7574768289923668,
      "num_tokens": 1808976064.0,
      "step": 25170
    },
    {
      "entropy": 0.8908263728022575,
      "epoch": 6.862905423821204,
      "grad_norm": 0.13036611676216125,
      "learning_rate": 6.106745699010518e-05,
      "loss": 1.3232,
      "mean_token_accuracy": 0.7502112478017807,
      "num_tokens": 1809687524.0,
      "step": 25180
    },
    {
      "entropy": 0.8970330774784088,
      "epoch": 6.865630962115017,
      "grad_norm": 0.129790797829628,
      "learning_rate": 6.103815613694584e-05,
      "loss": 1.3266,
      "mean_token_accuracy": 0.7507972061634064,
      "num_tokens": 1810411749.0,
      "step": 25190
    },
    {
      "entropy": 0.8960664182901382,
      "epoch": 6.868356500408831,
      "grad_norm": 0.13538548350334167,
      "learning_rate": 6.100885129835575e-05,
      "loss": 1.3263,
      "mean_token_accuracy": 0.7504530757665634,
      "num_tokens": 1811126021.0,
      "step": 25200
    },
    {
      "entropy": 0.898378475010395,
      "epoch": 6.871082038702644,
      "grad_norm": 0.13220278918743134,
      "learning_rate": 6.09795424849157e-05,
      "loss": 1.3228,
      "mean_token_accuracy": 0.7522737562656403,
      "num_tokens": 1811860168.0,
      "step": 25210
    },
    {
      "entropy": 0.8975155144929886,
      "epoch": 6.873807576996457,
      "grad_norm": 0.12303317338228226,
      "learning_rate": 6.095022970720793e-05,
      "loss": 1.3164,
      "mean_token_accuracy": 0.7558909565210342,
      "num_tokens": 1812600662.0,
      "step": 25220
    },
    {
      "entropy": 0.8922346726059913,
      "epoch": 6.87653311529027,
      "grad_norm": 0.1409824788570404,
      "learning_rate": 6.092091297581608e-05,
      "loss": 1.3189,
      "mean_token_accuracy": 0.7525821834802627,
      "num_tokens": 1813316952.0,
      "step": 25230
    },
    {
      "entropy": 0.8965624541044235,
      "epoch": 6.879258653584083,
      "grad_norm": 0.12102717906236649,
      "learning_rate": 6.089159230132524e-05,
      "loss": 1.3306,
      "mean_token_accuracy": 0.7498206123709679,
      "num_tokens": 1814030467.0,
      "step": 25240
    },
    {
      "entropy": 0.893408976495266,
      "epoch": 6.881984191877896,
      "grad_norm": 0.1330675482749939,
      "learning_rate": 6.086226769432193e-05,
      "loss": 1.3217,
      "mean_token_accuracy": 0.7534523352980613,
      "num_tokens": 1814755046.0,
      "step": 25250
    },
    {
      "entropy": 0.8845838963985443,
      "epoch": 6.884709730171709,
      "grad_norm": 0.12887269258499146,
      "learning_rate": 6.083293916539406e-05,
      "loss": 1.3155,
      "mean_token_accuracy": 0.7518935963511467,
      "num_tokens": 1815466463.0,
      "step": 25260
    },
    {
      "entropy": 0.8824153304100036,
      "epoch": 6.887435268465522,
      "grad_norm": 0.13421593606472015,
      "learning_rate": 6.0803606725130965e-05,
      "loss": 1.3119,
      "mean_token_accuracy": 0.7580133184790612,
      "num_tokens": 1816185402.0,
      "step": 25270
    },
    {
      "entropy": 0.8832359358668327,
      "epoch": 6.890160806759335,
      "grad_norm": 0.12394164502620697,
      "learning_rate": 6.077427038412344e-05,
      "loss": 1.3097,
      "mean_token_accuracy": 0.7570543676614762,
      "num_tokens": 1816905596.0,
      "step": 25280
    },
    {
      "entropy": 0.884107905626297,
      "epoch": 6.892886345053148,
      "grad_norm": 0.13406658172607422,
      "learning_rate": 6.074493015296361e-05,
      "loss": 1.3114,
      "mean_token_accuracy": 0.7576464235782623,
      "num_tokens": 1817640103.0,
      "step": 25290
    },
    {
      "entropy": 0.8825078994035721,
      "epoch": 6.895611883346961,
      "grad_norm": 0.14242766797542572,
      "learning_rate": 6.0715586042245043e-05,
      "loss": 1.3143,
      "mean_token_accuracy": 0.7528106242418289,
      "num_tokens": 1818339337.0,
      "step": 25300
    },
    {
      "entropy": 0.8900894001126289,
      "epoch": 6.898337421640774,
      "grad_norm": 0.1387607753276825,
      "learning_rate": 6.068623806256274e-05,
      "loss": 1.3203,
      "mean_token_accuracy": 0.7529940992593765,
      "num_tokens": 1819054669.0,
      "step": 25310
    },
    {
      "entropy": 0.88262540102005,
      "epoch": 6.901062959934587,
      "grad_norm": 0.12354110181331635,
      "learning_rate": 6.065688622451304e-05,
      "loss": 1.3113,
      "mean_token_accuracy": 0.7585963323712349,
      "num_tokens": 1819784957.0,
      "step": 25320
    },
    {
      "entropy": 0.8864086881279946,
      "epoch": 6.9037884982284,
      "grad_norm": 0.14237426221370697,
      "learning_rate": 6.06275305386937e-05,
      "loss": 1.3144,
      "mean_token_accuracy": 0.7556328341364861,
      "num_tokens": 1820507113.0,
      "step": 25330
    },
    {
      "entropy": 0.8944937258958816,
      "epoch": 6.906514036522213,
      "grad_norm": 0.13102467358112335,
      "learning_rate": 6.0598171015703895e-05,
      "loss": 1.3242,
      "mean_token_accuracy": 0.75197584182024,
      "num_tokens": 1821219311.0,
      "step": 25340
    },
    {
      "entropy": 0.8945687383413314,
      "epoch": 6.909239574816026,
      "grad_norm": 0.1266109198331833,
      "learning_rate": 6.056880766614412e-05,
      "loss": 1.3243,
      "mean_token_accuracy": 0.7514003947377205,
      "num_tokens": 1821932908.0,
      "step": 25350
    },
    {
      "entropy": 0.8875587612390519,
      "epoch": 6.911965113109839,
      "grad_norm": 0.12885835766792297,
      "learning_rate": 6.053944050061633e-05,
      "loss": 1.3169,
      "mean_token_accuracy": 0.7505095466971398,
      "num_tokens": 1822634108.0,
      "step": 25360
    },
    {
      "entropy": 0.8883148550987243,
      "epoch": 6.914690651403652,
      "grad_norm": 0.1644582897424698,
      "learning_rate": 6.051006952972378e-05,
      "loss": 1.3136,
      "mean_token_accuracy": 0.7556955307722092,
      "num_tokens": 1823355702.0,
      "step": 25370
    },
    {
      "entropy": 0.8857153192162514,
      "epoch": 6.917416189697465,
      "grad_norm": 0.16814067959785461,
      "learning_rate": 6.048069476407117e-05,
      "loss": 1.3129,
      "mean_token_accuracy": 0.7535390347242356,
      "num_tokens": 1824060649.0,
      "step": 25380
    },
    {
      "entropy": 0.888411970436573,
      "epoch": 6.920141727991278,
      "grad_norm": 0.13149769604206085,
      "learning_rate": 6.045131621426452e-05,
      "loss": 1.3127,
      "mean_token_accuracy": 0.7515056937932968,
      "num_tokens": 1824773374.0,
      "step": 25390
    },
    {
      "entropy": 0.8872165113687516,
      "epoch": 6.922867266285091,
      "grad_norm": 0.1266518384218216,
      "learning_rate": 6.042193389091124e-05,
      "loss": 1.3186,
      "mean_token_accuracy": 0.7529063448309898,
      "num_tokens": 1825489650.0,
      "step": 25400
    },
    {
      "entropy": 0.8882512658834457,
      "epoch": 6.925592804578905,
      "grad_norm": 0.14183460175991058,
      "learning_rate": 6.0392547804620093e-05,
      "loss": 1.3135,
      "mean_token_accuracy": 0.7538353443145752,
      "num_tokens": 1826212918.0,
      "step": 25410
    },
    {
      "entropy": 0.8908900067210197,
      "epoch": 6.928318342872718,
      "grad_norm": 0.14262664318084717,
      "learning_rate": 6.036315796600119e-05,
      "loss": 1.3196,
      "mean_token_accuracy": 0.7542007371783257,
      "num_tokens": 1826928865.0,
      "step": 25420
    },
    {
      "entropy": 0.8866520047187805,
      "epoch": 6.931043881166531,
      "grad_norm": 0.12231040745973587,
      "learning_rate": 6.033376438566605e-05,
      "loss": 1.3015,
      "mean_token_accuracy": 0.755017101764679,
      "num_tokens": 1827654261.0,
      "step": 25430
    },
    {
      "entropy": 0.8829600974917412,
      "epoch": 6.933769419460344,
      "grad_norm": 0.13049496710300446,
      "learning_rate": 6.030436707422745e-05,
      "loss": 1.3072,
      "mean_token_accuracy": 0.7576208561658859,
      "num_tokens": 1828376179.0,
      "step": 25440
    },
    {
      "entropy": 0.8899279996752739,
      "epoch": 6.936494957754157,
      "grad_norm": 0.13958017528057098,
      "learning_rate": 6.027496604229961e-05,
      "loss": 1.3224,
      "mean_token_accuracy": 0.7475732579827309,
      "num_tokens": 1829068793.0,
      "step": 25450
    },
    {
      "entropy": 0.8850479483604431,
      "epoch": 6.9392204960479695,
      "grad_norm": 0.13414804637432098,
      "learning_rate": 6.0245561300498034e-05,
      "loss": 1.3207,
      "mean_token_accuracy": 0.7567609786987305,
      "num_tokens": 1829789136.0,
      "step": 25460
    },
    {
      "entropy": 0.8858461394906044,
      "epoch": 6.9419460343417825,
      "grad_norm": 0.12835468351840973,
      "learning_rate": 6.021615285943958e-05,
      "loss": 1.3133,
      "mean_token_accuracy": 0.7571169376373291,
      "num_tokens": 1830510355.0,
      "step": 25470
    },
    {
      "entropy": 0.8861836612224578,
      "epoch": 6.944671572635595,
      "grad_norm": 0.13081277906894684,
      "learning_rate": 6.018674072974244e-05,
      "loss": 1.308,
      "mean_token_accuracy": 0.7534803554415703,
      "num_tokens": 1831233640.0,
      "step": 25480
    },
    {
      "entropy": 0.8840690895915031,
      "epoch": 6.947397110929408,
      "grad_norm": 0.1314501315355301,
      "learning_rate": 6.015732492202617e-05,
      "loss": 1.3125,
      "mean_token_accuracy": 0.7585026815533638,
      "num_tokens": 1831975632.0,
      "step": 25490
    },
    {
      "entropy": 0.8896793007850647,
      "epoch": 6.950122649223221,
      "grad_norm": 0.14163920283317566,
      "learning_rate": 6.012790544691159e-05,
      "loss": 1.3101,
      "mean_token_accuracy": 0.7513944432139397,
      "num_tokens": 1832690253.0,
      "step": 25500
    },
    {
      "entropy": 0.8853412166237831,
      "epoch": 6.952848187517034,
      "grad_norm": 0.13058798015117645,
      "learning_rate": 6.0098482315020896e-05,
      "loss": 1.3069,
      "mean_token_accuracy": 0.7581047922372818,
      "num_tokens": 1833423579.0,
      "step": 25510
    },
    {
      "entropy": 0.8871693447232246,
      "epoch": 6.955573725810847,
      "grad_norm": 0.12563154101371765,
      "learning_rate": 6.0069055536977595e-05,
      "loss": 1.3158,
      "mean_token_accuracy": 0.7564632892608643,
      "num_tokens": 1834134410.0,
      "step": 25520
    },
    {
      "entropy": 0.882788622379303,
      "epoch": 6.958299264104661,
      "grad_norm": 0.1307438313961029,
      "learning_rate": 6.0039625123406495e-05,
      "loss": 1.3099,
      "mean_token_accuracy": 0.7566651105880737,
      "num_tokens": 1834858003.0,
      "step": 25530
    },
    {
      "entropy": 0.8913815826177597,
      "epoch": 6.961024802398474,
      "grad_norm": 0.13012726604938507,
      "learning_rate": 6.001019108493372e-05,
      "loss": 1.3145,
      "mean_token_accuracy": 0.7526103541254997,
      "num_tokens": 1835573847.0,
      "step": 25540
    },
    {
      "entropy": 0.8850127249956131,
      "epoch": 6.963750340692287,
      "grad_norm": 0.12452500313520432,
      "learning_rate": 5.998075343218672e-05,
      "loss": 1.3105,
      "mean_token_accuracy": 0.7558508038520813,
      "num_tokens": 1836291597.0,
      "step": 25550
    },
    {
      "entropy": 0.8863060176372528,
      "epoch": 6.9664758789861,
      "grad_norm": 0.12626658380031586,
      "learning_rate": 5.995131217579423e-05,
      "loss": 1.3131,
      "mean_token_accuracy": 0.7589380100369454,
      "num_tokens": 1837028787.0,
      "step": 25560
    },
    {
      "entropy": 0.8877708554267884,
      "epoch": 6.969201417279913,
      "grad_norm": 0.13097473978996277,
      "learning_rate": 5.9921867326386294e-05,
      "loss": 1.3134,
      "mean_token_accuracy": 0.7542364045977592,
      "num_tokens": 1837757482.0,
      "step": 25570
    },
    {
      "entropy": 0.894642573595047,
      "epoch": 6.971926955573726,
      "grad_norm": 0.1353239119052887,
      "learning_rate": 5.989241889459427e-05,
      "loss": 1.3181,
      "mean_token_accuracy": 0.7549946740269661,
      "num_tokens": 1838492788.0,
      "step": 25580
    },
    {
      "entropy": 0.8870107933878899,
      "epoch": 6.974652493867539,
      "grad_norm": 0.12750038504600525,
      "learning_rate": 5.986296689105076e-05,
      "loss": 1.3143,
      "mean_token_accuracy": 0.7520146772265435,
      "num_tokens": 1839205036.0,
      "step": 25590
    },
    {
      "entropy": 0.8909857362508774,
      "epoch": 6.977378032161352,
      "grad_norm": 0.12656597793102264,
      "learning_rate": 5.9833511326389725e-05,
      "loss": 1.3124,
      "mean_token_accuracy": 0.756410850584507,
      "num_tokens": 1839936330.0,
      "step": 25600
    },
    {
      "entropy": 0.8968017011880874,
      "epoch": 6.980103570455165,
      "grad_norm": 0.12405797094106674,
      "learning_rate": 5.9804052211246345e-05,
      "loss": 1.3199,
      "mean_token_accuracy": 0.7516512796282768,
      "num_tokens": 1840651268.0,
      "step": 25610
    },
    {
      "entropy": 0.8871091291308403,
      "epoch": 6.9828291087489776,
      "grad_norm": 0.13120581209659576,
      "learning_rate": 5.977458955625712e-05,
      "loss": 1.3099,
      "mean_token_accuracy": 0.7555394068360328,
      "num_tokens": 1841387356.0,
      "step": 25620
    },
    {
      "entropy": 0.8886284098029137,
      "epoch": 6.985554647042791,
      "grad_norm": 0.13083769381046295,
      "learning_rate": 5.9745123372059853e-05,
      "loss": 1.3123,
      "mean_token_accuracy": 0.7536327734589576,
      "num_tokens": 1842115008.0,
      "step": 25630
    },
    {
      "entropy": 0.8853094175457954,
      "epoch": 6.988280185336604,
      "grad_norm": 0.1331971287727356,
      "learning_rate": 5.9715653669293545e-05,
      "loss": 1.3047,
      "mean_token_accuracy": 0.7562429383397102,
      "num_tokens": 1842840643.0,
      "step": 25640
    },
    {
      "entropy": 0.8968934863805771,
      "epoch": 6.991005723630417,
      "grad_norm": 0.1355733871459961,
      "learning_rate": 5.968618045859853e-05,
      "loss": 1.3269,
      "mean_token_accuracy": 0.7489420115947724,
      "num_tokens": 1843545441.0,
      "step": 25650
    },
    {
      "entropy": 0.8933965265750885,
      "epoch": 6.99373126192423,
      "grad_norm": 0.12968383729457855,
      "learning_rate": 5.965670375061638e-05,
      "loss": 1.3147,
      "mean_token_accuracy": 0.7529557570815086,
      "num_tokens": 1844262644.0,
      "step": 25660
    },
    {
      "entropy": 0.8912503585219383,
      "epoch": 6.996456800218043,
      "grad_norm": 0.12128704786300659,
      "learning_rate": 5.9627223555989974e-05,
      "loss": 1.3192,
      "mean_token_accuracy": 0.7518787369132042,
      "num_tokens": 1844978661.0,
      "step": 25670
    },
    {
      "entropy": 0.8842117726802826,
      "epoch": 6.999182338511856,
      "grad_norm": 0.12942983210086823,
      "learning_rate": 5.959773988536337e-05,
      "loss": 1.3122,
      "mean_token_accuracy": 0.753253199160099,
      "num_tokens": 1845689292.0,
      "step": 25680
    },
    {
      "entropy": 0.8827182292938233,
      "epoch": 7.001907876805669,
      "grad_norm": 0.12594494223594666,
      "learning_rate": 5.9568252749381946e-05,
      "loss": 1.3043,
      "mean_token_accuracy": 0.7580388650298119,
      "num_tokens": 1846409716.0,
      "step": 25690
    },
    {
      "entropy": 0.87383753657341,
      "epoch": 7.004633415099482,
      "grad_norm": 0.1468559056520462,
      "learning_rate": 5.953876215869231e-05,
      "loss": 1.2991,
      "mean_token_accuracy": 0.7622627973556518,
      "num_tokens": 1847148069.0,
      "step": 25700
    },
    {
      "entropy": 0.8747234106063843,
      "epoch": 7.007358953393295,
      "grad_norm": 0.14879809319972992,
      "learning_rate": 5.950926812394232e-05,
      "loss": 1.299,
      "mean_token_accuracy": 0.7560095652937889,
      "num_tokens": 1847870227.0,
      "step": 25710
    },
    {
      "entropy": 0.874944044649601,
      "epoch": 7.010084491687108,
      "grad_norm": 0.14737948775291443,
      "learning_rate": 5.947977065578108e-05,
      "loss": 1.3028,
      "mean_token_accuracy": 0.7561734259128571,
      "num_tokens": 1848587072.0,
      "step": 25720
    },
    {
      "entropy": 0.8718336135149002,
      "epoch": 7.012810029980921,
      "grad_norm": 0.14516662061214447,
      "learning_rate": 5.945026976485891e-05,
      "loss": 1.2919,
      "mean_token_accuracy": 0.7556194886565208,
      "num_tokens": 1849303510.0,
      "step": 25730
    },
    {
      "entropy": 0.8663932263851166,
      "epoch": 7.015535568274735,
      "grad_norm": 0.13793468475341797,
      "learning_rate": 5.9420765461827424e-05,
      "loss": 1.291,
      "mean_token_accuracy": 0.7601346105337143,
      "num_tokens": 1850029730.0,
      "step": 25740
    },
    {
      "entropy": 0.8823228001594543,
      "epoch": 7.018261106568548,
      "grad_norm": 0.15650111436843872,
      "learning_rate": 5.939125775733939e-05,
      "loss": 1.3005,
      "mean_token_accuracy": 0.7535691395401954,
      "num_tokens": 1850744487.0,
      "step": 25750
    },
    {
      "entropy": 0.8704546928405762,
      "epoch": 7.020986644862361,
      "grad_norm": 0.13706140220165253,
      "learning_rate": 5.936174666204887e-05,
      "loss": 1.29,
      "mean_token_accuracy": 0.7583501160144805,
      "num_tokens": 1851474367.0,
      "step": 25760
    },
    {
      "entropy": 0.8767647996544838,
      "epoch": 7.0237121831561735,
      "grad_norm": 0.15674827992916107,
      "learning_rate": 5.933223218661111e-05,
      "loss": 1.2937,
      "mean_token_accuracy": 0.7525588110089302,
      "num_tokens": 1852182499.0,
      "step": 25770
    },
    {
      "entropy": 0.8705548405647278,
      "epoch": 7.0264377214499865,
      "grad_norm": 0.14327582716941833,
      "learning_rate": 5.930271434168261e-05,
      "loss": 1.2879,
      "mean_token_accuracy": 0.7598174691200257,
      "num_tokens": 1852911780.0,
      "step": 25780
    },
    {
      "entropy": 0.8777773588895798,
      "epoch": 7.029163259743799,
      "grad_norm": 0.16943815350532532,
      "learning_rate": 5.9273193137921034e-05,
      "loss": 1.3014,
      "mean_token_accuracy": 0.7552996784448623,
      "num_tokens": 1853624809.0,
      "step": 25790
    },
    {
      "entropy": 0.8729385986924172,
      "epoch": 7.031888798037612,
      "grad_norm": 0.14058621227741241,
      "learning_rate": 5.9243668585985324e-05,
      "loss": 1.2974,
      "mean_token_accuracy": 0.7559783667325973,
      "num_tokens": 1854332134.0,
      "step": 25800
    },
    {
      "entropy": 0.8762853026390076,
      "epoch": 7.034614336331425,
      "grad_norm": 0.15448784828186035,
      "learning_rate": 5.921414069653559e-05,
      "loss": 1.2886,
      "mean_token_accuracy": 0.7573204502463341,
      "num_tokens": 1855055077.0,
      "step": 25810
    },
    {
      "entropy": 0.8673344030976295,
      "epoch": 7.037339874625238,
      "grad_norm": 0.14283548295497894,
      "learning_rate": 5.918460948023314e-05,
      "loss": 1.2904,
      "mean_token_accuracy": 0.761931137740612,
      "num_tokens": 1855789834.0,
      "step": 25820
    },
    {
      "entropy": 0.8762605920433998,
      "epoch": 7.040065412919051,
      "grad_norm": 0.14536963403224945,
      "learning_rate": 5.915507494774051e-05,
      "loss": 1.307,
      "mean_token_accuracy": 0.7530893743038177,
      "num_tokens": 1856499524.0,
      "step": 25830
    },
    {
      "entropy": 0.8695098787546158,
      "epoch": 7.042790951212864,
      "grad_norm": 0.14235912263393402,
      "learning_rate": 5.912553710972143e-05,
      "loss": 1.2879,
      "mean_token_accuracy": 0.7548491284251213,
      "num_tokens": 1857203651.0,
      "step": 25840
    },
    {
      "entropy": 0.8765233427286148,
      "epoch": 7.045516489506678,
      "grad_norm": 0.15693019330501556,
      "learning_rate": 5.909599597684081e-05,
      "loss": 1.29,
      "mean_token_accuracy": 0.7588856726884842,
      "num_tokens": 1857930689.0,
      "step": 25850
    },
    {
      "entropy": 0.873693810403347,
      "epoch": 7.048242027800491,
      "grad_norm": 0.1523929387331009,
      "learning_rate": 5.906645155976476e-05,
      "loss": 1.2999,
      "mean_token_accuracy": 0.7557924911379814,
      "num_tokens": 1858652937.0,
      "step": 25860
    },
    {
      "entropy": 0.8805491030216217,
      "epoch": 7.050967566094304,
      "grad_norm": 0.13906849920749664,
      "learning_rate": 5.9036903869160564e-05,
      "loss": 1.3049,
      "mean_token_accuracy": 0.7498277887701988,
      "num_tokens": 1859351039.0,
      "step": 25870
    },
    {
      "entropy": 0.877228170633316,
      "epoch": 7.053693104388117,
      "grad_norm": 0.14655278623104095,
      "learning_rate": 5.90073529156967e-05,
      "loss": 1.2929,
      "mean_token_accuracy": 0.7556937620043754,
      "num_tokens": 1860073562.0,
      "step": 25880
    },
    {
      "entropy": 0.8722148582339286,
      "epoch": 7.05641864268193,
      "grad_norm": 0.14586222171783447,
      "learning_rate": 5.897779871004281e-05,
      "loss": 1.2926,
      "mean_token_accuracy": 0.7574687734246254,
      "num_tokens": 1860806603.0,
      "step": 25890
    },
    {
      "entropy": 0.8758646205067635,
      "epoch": 7.059144180975743,
      "grad_norm": 0.14526154100894928,
      "learning_rate": 5.894824126286974e-05,
      "loss": 1.3014,
      "mean_token_accuracy": 0.7530226081609726,
      "num_tokens": 1861517748.0,
      "step": 25900
    },
    {
      "entropy": 0.8793660134077073,
      "epoch": 7.061869719269556,
      "grad_norm": 0.14603964984416962,
      "learning_rate": 5.8918680584849464e-05,
      "loss": 1.2997,
      "mean_token_accuracy": 0.7553532078862191,
      "num_tokens": 1862239234.0,
      "step": 25910
    },
    {
      "entropy": 0.8771114945411682,
      "epoch": 7.064595257563369,
      "grad_norm": 0.15067288279533386,
      "learning_rate": 5.8889116686655155e-05,
      "loss": 1.3032,
      "mean_token_accuracy": 0.756546376645565,
      "num_tokens": 1862954604.0,
      "step": 25920
    },
    {
      "entropy": 0.8750939950346946,
      "epoch": 7.0673207958571815,
      "grad_norm": 0.14782260358333588,
      "learning_rate": 5.885954957896115e-05,
      "loss": 1.2938,
      "mean_token_accuracy": 0.7629788026213646,
      "num_tokens": 1863688097.0,
      "step": 25930
    },
    {
      "entropy": 0.8741045355796814,
      "epoch": 7.0700463341509945,
      "grad_norm": 0.14869625866413116,
      "learning_rate": 5.88299792724429e-05,
      "loss": 1.2993,
      "mean_token_accuracy": 0.759143902361393,
      "num_tokens": 1864416271.0,
      "step": 25940
    },
    {
      "entropy": 0.8720627635717392,
      "epoch": 7.072771872444807,
      "grad_norm": 0.14257268607616425,
      "learning_rate": 5.880040577777708e-05,
      "loss": 1.3009,
      "mean_token_accuracy": 0.7576109483838082,
      "num_tokens": 1865134311.0,
      "step": 25950
    },
    {
      "entropy": 0.8788901478052139,
      "epoch": 7.075497410738621,
      "grad_norm": 0.15309759974479675,
      "learning_rate": 5.877082910564146e-05,
      "loss": 1.3063,
      "mean_token_accuracy": 0.7543908953666687,
      "num_tokens": 1865845673.0,
      "step": 25960
    },
    {
      "entropy": 0.8736586809158325,
      "epoch": 7.078222949032434,
      "grad_norm": 0.15360787510871887,
      "learning_rate": 5.874124926671498e-05,
      "loss": 1.3032,
      "mean_token_accuracy": 0.756864820420742,
      "num_tokens": 1866558735.0,
      "step": 25970
    },
    {
      "entropy": 0.8797694593667984,
      "epoch": 7.080948487326247,
      "grad_norm": 0.1488131731748581,
      "learning_rate": 5.871166627167772e-05,
      "loss": 1.2994,
      "mean_token_accuracy": 0.7545985385775567,
      "num_tokens": 1867273078.0,
      "step": 25980
    },
    {
      "entropy": 0.8750902444124222,
      "epoch": 7.08367402562006,
      "grad_norm": 0.1553211510181427,
      "learning_rate": 5.8682080131210916e-05,
      "loss": 1.299,
      "mean_token_accuracy": 0.7520381003618241,
      "num_tokens": 1867972370.0,
      "step": 25990
    },
    {
      "entropy": 0.8790387660264969,
      "epoch": 7.086399563913873,
      "grad_norm": 0.14957445859909058,
      "learning_rate": 5.865249085599689e-05,
      "loss": 1.2963,
      "mean_token_accuracy": 0.7567360460758209,
      "num_tokens": 1868707356.0,
      "step": 26000
    },
    {
      "entropy": 0.8829400196671486,
      "epoch": 7.089125102207686,
      "grad_norm": 0.16321925818920135,
      "learning_rate": 5.862289845671917e-05,
      "loss": 1.2954,
      "mean_token_accuracy": 0.7559129923582077,
      "num_tokens": 1869432113.0,
      "step": 26010
    },
    {
      "entropy": 0.8690956473350525,
      "epoch": 7.091850640501499,
      "grad_norm": 0.1361200362443924,
      "learning_rate": 5.8593302944062346e-05,
      "loss": 1.2987,
      "mean_token_accuracy": 0.7589480504393578,
      "num_tokens": 1870150186.0,
      "step": 26020
    },
    {
      "entropy": 0.8744716703891754,
      "epoch": 7.094576178795312,
      "grad_norm": 0.14653931558132172,
      "learning_rate": 5.856370432871216e-05,
      "loss": 1.3006,
      "mean_token_accuracy": 0.7579378753900528,
      "num_tokens": 1870872565.0,
      "step": 26030
    },
    {
      "entropy": 0.8781326964497567,
      "epoch": 7.097301717089125,
      "grad_norm": 0.1471918225288391,
      "learning_rate": 5.853410262135548e-05,
      "loss": 1.3025,
      "mean_token_accuracy": 0.7564917609095574,
      "num_tokens": 1871590788.0,
      "step": 26040
    },
    {
      "entropy": 0.877470214664936,
      "epoch": 7.100027255382938,
      "grad_norm": 0.14674139022827148,
      "learning_rate": 5.850449783268029e-05,
      "loss": 1.3018,
      "mean_token_accuracy": 0.754204623401165,
      "num_tokens": 1872306796.0,
      "step": 26050
    },
    {
      "entropy": 0.8754927709698677,
      "epoch": 7.102752793676752,
      "grad_norm": 0.1696721315383911,
      "learning_rate": 5.847488997337566e-05,
      "loss": 1.3006,
      "mean_token_accuracy": 0.7581387490034104,
      "num_tokens": 1873028300.0,
      "step": 26060
    },
    {
      "entropy": 0.8749557644128799,
      "epoch": 7.1054783319705646,
      "grad_norm": 0.1399001181125641,
      "learning_rate": 5.8445279054131785e-05,
      "loss": 1.2996,
      "mean_token_accuracy": 0.7563225671648979,
      "num_tokens": 1873749232.0,
      "step": 26070
    },
    {
      "entropy": 0.8754014566540718,
      "epoch": 7.1082038702643775,
      "grad_norm": 0.13720010221004486,
      "learning_rate": 5.841566508563999e-05,
      "loss": 1.294,
      "mean_token_accuracy": 0.7545505180954933,
      "num_tokens": 1874462772.0,
      "step": 26080
    },
    {
      "entropy": 0.8690753966569901,
      "epoch": 7.1109294085581904,
      "grad_norm": 0.15143989026546478,
      "learning_rate": 5.838604807859267e-05,
      "loss": 1.2934,
      "mean_token_accuracy": 0.7571880966424942,
      "num_tokens": 1875193775.0,
      "step": 26090
    },
    {
      "entropy": 0.8689525797963142,
      "epoch": 7.113654946852003,
      "grad_norm": 0.14893269538879395,
      "learning_rate": 5.835642804368333e-05,
      "loss": 1.2914,
      "mean_token_accuracy": 0.7562677934765816,
      "num_tokens": 1875902636.0,
      "step": 26100
    },
    {
      "entropy": 0.8736646234989166,
      "epoch": 7.116380485145816,
      "grad_norm": 0.1414748728275299,
      "learning_rate": 5.8326804991606546e-05,
      "loss": 1.295,
      "mean_token_accuracy": 0.757441146671772,
      "num_tokens": 1876626275.0,
      "step": 26110
    },
    {
      "entropy": 0.8731375485658646,
      "epoch": 7.119106023439629,
      "grad_norm": 0.14848846197128296,
      "learning_rate": 5.829717893305801e-05,
      "loss": 1.2884,
      "mean_token_accuracy": 0.7576317623257637,
      "num_tokens": 1877347469.0,
      "step": 26120
    },
    {
      "entropy": 0.8723952531814575,
      "epoch": 7.121831561733442,
      "grad_norm": 0.1503380537033081,
      "learning_rate": 5.8267549878734495e-05,
      "loss": 1.305,
      "mean_token_accuracy": 0.7554426342248917,
      "num_tokens": 1878061805.0,
      "step": 26130
    },
    {
      "entropy": 0.8748130679130555,
      "epoch": 7.124557100027255,
      "grad_norm": 0.1450539529323578,
      "learning_rate": 5.823791783933386e-05,
      "loss": 1.2991,
      "mean_token_accuracy": 0.7596857830882072,
      "num_tokens": 1878799450.0,
      "step": 26140
    },
    {
      "entropy": 0.869141885638237,
      "epoch": 7.127282638321068,
      "grad_norm": 0.15486854314804077,
      "learning_rate": 5.820828282555502e-05,
      "loss": 1.2931,
      "mean_token_accuracy": 0.7584925502538681,
      "num_tokens": 1879519878.0,
      "step": 26150
    },
    {
      "entropy": 0.8671886593103408,
      "epoch": 7.130008176614881,
      "grad_norm": 0.15078620612621307,
      "learning_rate": 5.817864484809797e-05,
      "loss": 1.285,
      "mean_token_accuracy": 0.7625649064779282,
      "num_tokens": 1880255117.0,
      "step": 26160
    },
    {
      "entropy": 0.876307912170887,
      "epoch": 7.132733714908694,
      "grad_norm": 0.15839825570583344,
      "learning_rate": 5.81490039176638e-05,
      "loss": 1.3062,
      "mean_token_accuracy": 0.7533732756972313,
      "num_tokens": 1880960324.0,
      "step": 26170
    },
    {
      "entropy": 0.8709997043013573,
      "epoch": 7.135459253202508,
      "grad_norm": 0.13962945342063904,
      "learning_rate": 5.811936004495465e-05,
      "loss": 1.3044,
      "mean_token_accuracy": 0.7553649961948394,
      "num_tokens": 1881663381.0,
      "step": 26180
    },
    {
      "entropy": 0.8603877604007721,
      "epoch": 7.138184791496321,
      "grad_norm": 0.1384153664112091,
      "learning_rate": 5.8089713240673695e-05,
      "loss": 1.292,
      "mean_token_accuracy": 0.7593579024076462,
      "num_tokens": 1882383259.0,
      "step": 26190
    },
    {
      "entropy": 0.8779341742396355,
      "epoch": 7.140910329790134,
      "grad_norm": 0.1410437822341919,
      "learning_rate": 5.806006351552523e-05,
      "loss": 1.2977,
      "mean_token_accuracy": 0.7531718775629997,
      "num_tokens": 1883096871.0,
      "step": 26200
    },
    {
      "entropy": 0.8759750500321388,
      "epoch": 7.143635868083947,
      "grad_norm": 0.1437990814447403,
      "learning_rate": 5.8030410880214535e-05,
      "loss": 1.2919,
      "mean_token_accuracy": 0.7562564298510551,
      "num_tokens": 1883817598.0,
      "step": 26210
    },
    {
      "entropy": 0.870393043756485,
      "epoch": 7.14636140637776,
      "grad_norm": 0.14680783450603485,
      "learning_rate": 5.8000755345448e-05,
      "loss": 1.2938,
      "mean_token_accuracy": 0.7575946375727654,
      "num_tokens": 1884540367.0,
      "step": 26220
    },
    {
      "entropy": 0.8708333566784858,
      "epoch": 7.149086944671573,
      "grad_norm": 0.1478429138660431,
      "learning_rate": 5.797109692193302e-05,
      "loss": 1.2992,
      "mean_token_accuracy": 0.758022865653038,
      "num_tokens": 1885258891.0,
      "step": 26230
    },
    {
      "entropy": 0.8699542701244354,
      "epoch": 7.1518124829653855,
      "grad_norm": 0.24645479023456573,
      "learning_rate": 5.794143562037806e-05,
      "loss": 1.2889,
      "mean_token_accuracy": 0.7612785547971725,
      "num_tokens": 1885982493.0,
      "step": 26240
    },
    {
      "entropy": 0.8763674259185791,
      "epoch": 7.1545380212591985,
      "grad_norm": 0.16096802055835724,
      "learning_rate": 5.79117714514926e-05,
      "loss": 1.3015,
      "mean_token_accuracy": 0.7544290259480476,
      "num_tokens": 1886692035.0,
      "step": 26250
    },
    {
      "entropy": 0.8811359405517578,
      "epoch": 7.157263559553011,
      "grad_norm": 0.15165232121944427,
      "learning_rate": 5.788210442598719e-05,
      "loss": 1.2984,
      "mean_token_accuracy": 0.7553988859057427,
      "num_tokens": 1887419643.0,
      "step": 26260
    },
    {
      "entropy": 0.8746663436293602,
      "epoch": 7.159989097846824,
      "grad_norm": 0.13912591338157654,
      "learning_rate": 5.785243455457338e-05,
      "loss": 1.2987,
      "mean_token_accuracy": 0.7545264646410942,
      "num_tokens": 1888143199.0,
      "step": 26270
    },
    {
      "entropy": 0.8752788513898849,
      "epoch": 7.162714636140638,
      "grad_norm": 0.17055462300777435,
      "learning_rate": 5.782276184796373e-05,
      "loss": 1.2935,
      "mean_token_accuracy": 0.7571158647537232,
      "num_tokens": 1888869073.0,
      "step": 26280
    },
    {
      "entropy": 0.8772252917289733,
      "epoch": 7.165440174434451,
      "grad_norm": 0.1450638771057129,
      "learning_rate": 5.779308631687191e-05,
      "loss": 1.297,
      "mean_token_accuracy": 0.7533574894070625,
      "num_tokens": 1889587743.0,
      "step": 26290
    },
    {
      "entropy": 0.8710338369011879,
      "epoch": 7.168165712728264,
      "grad_norm": 0.1593686193227768,
      "learning_rate": 5.776340797201252e-05,
      "loss": 1.2938,
      "mean_token_accuracy": 0.7594601094722748,
      "num_tokens": 1890317769.0,
      "step": 26300
    },
    {
      "entropy": 0.8735907509922981,
      "epoch": 7.170891251022077,
      "grad_norm": 0.14964617788791656,
      "learning_rate": 5.773372682410119e-05,
      "loss": 1.2931,
      "mean_token_accuracy": 0.7603439137339592,
      "num_tokens": 1891043657.0,
      "step": 26310
    },
    {
      "entropy": 0.8753811821341515,
      "epoch": 7.17361678931589,
      "grad_norm": 0.14793598651885986,
      "learning_rate": 5.770404288385459e-05,
      "loss": 1.293,
      "mean_token_accuracy": 0.7586567997932434,
      "num_tokens": 1891775050.0,
      "step": 26320
    },
    {
      "entropy": 0.8740849614143371,
      "epoch": 7.176342327609703,
      "grad_norm": 0.15123946964740753,
      "learning_rate": 5.767435616199041e-05,
      "loss": 1.2958,
      "mean_token_accuracy": 0.752157972753048,
      "num_tokens": 1892471265.0,
      "step": 26330
    },
    {
      "entropy": 0.8787199497222901,
      "epoch": 7.179067865903516,
      "grad_norm": 0.1432448923587799,
      "learning_rate": 5.7644666669227295e-05,
      "loss": 1.3021,
      "mean_token_accuracy": 0.7555218532681465,
      "num_tokens": 1893181058.0,
      "step": 26340
    },
    {
      "entropy": 0.8780265122652053,
      "epoch": 7.181793404197329,
      "grad_norm": 0.14211753010749817,
      "learning_rate": 5.761497441628494e-05,
      "loss": 1.3011,
      "mean_token_accuracy": 0.7550155445933342,
      "num_tokens": 1893907513.0,
      "step": 26350
    },
    {
      "entropy": 0.8750667542219162,
      "epoch": 7.184518942491142,
      "grad_norm": 0.16624762117862701,
      "learning_rate": 5.758527941388398e-05,
      "loss": 1.298,
      "mean_token_accuracy": 0.7573418155312538,
      "num_tokens": 1894631206.0,
      "step": 26360
    },
    {
      "entropy": 0.864818049967289,
      "epoch": 7.187244480784955,
      "grad_norm": 0.16290631890296936,
      "learning_rate": 5.755558167274613e-05,
      "loss": 1.289,
      "mean_token_accuracy": 0.7592792555689811,
      "num_tokens": 1895357144.0,
      "step": 26370
    },
    {
      "entropy": 0.8728147938847541,
      "epoch": 7.189970019078768,
      "grad_norm": 0.15238237380981445,
      "learning_rate": 5.752588120359399e-05,
      "loss": 1.2917,
      "mean_token_accuracy": 0.7546495750546456,
      "num_tokens": 1896070079.0,
      "step": 26380
    },
    {
      "entropy": 0.8765429362654686,
      "epoch": 7.1926955573725815,
      "grad_norm": 0.15025635063648224,
      "learning_rate": 5.749617801715121e-05,
      "loss": 1.3078,
      "mean_token_accuracy": 0.7570309311151504,
      "num_tokens": 1896793112.0,
      "step": 26390
    },
    {
      "entropy": 0.8747450307011604,
      "epoch": 7.195421095666394,
      "grad_norm": 0.14232859015464783,
      "learning_rate": 5.746647212414242e-05,
      "loss": 1.3013,
      "mean_token_accuracy": 0.7579892069101334,
      "num_tokens": 1897510056.0,
      "step": 26400
    },
    {
      "entropy": 0.8638418808579444,
      "epoch": 7.198146633960207,
      "grad_norm": 0.15131346881389618,
      "learning_rate": 5.74367635352932e-05,
      "loss": 1.2939,
      "mean_token_accuracy": 0.7587176471948623,
      "num_tokens": 1898214594.0,
      "step": 26410
    },
    {
      "entropy": 0.8682623416185379,
      "epoch": 7.20087217225402,
      "grad_norm": 0.15739648044109344,
      "learning_rate": 5.740705226133013e-05,
      "loss": 1.2891,
      "mean_token_accuracy": 0.7574314340949059,
      "num_tokens": 1898932898.0,
      "step": 26420
    },
    {
      "entropy": 0.8732482492923737,
      "epoch": 7.203597710547833,
      "grad_norm": 0.15174973011016846,
      "learning_rate": 5.7377338312980735e-05,
      "loss": 1.3013,
      "mean_token_accuracy": 0.7557144984602928,
      "num_tokens": 1899643041.0,
      "step": 26430
    },
    {
      "entropy": 0.8734035849571228,
      "epoch": 7.206323248841646,
      "grad_norm": 0.14822854101657867,
      "learning_rate": 5.734762170097352e-05,
      "loss": 1.3017,
      "mean_token_accuracy": 0.75667884349823,
      "num_tokens": 1900357545.0,
      "step": 26440
    },
    {
      "entropy": 0.8675457984209061,
      "epoch": 7.209048787135459,
      "grad_norm": 0.1542288213968277,
      "learning_rate": 5.731790243603794e-05,
      "loss": 1.2975,
      "mean_token_accuracy": 0.7582554280757904,
      "num_tokens": 1901064344.0,
      "step": 26450
    },
    {
      "entropy": 0.8695284485816955,
      "epoch": 7.211774325429272,
      "grad_norm": 0.15805630385875702,
      "learning_rate": 5.7288180528904445e-05,
      "loss": 1.2908,
      "mean_token_accuracy": 0.7603192314505577,
      "num_tokens": 1901782335.0,
      "step": 26460
    },
    {
      "entropy": 0.8807531267404556,
      "epoch": 7.214499863723085,
      "grad_norm": 0.13325440883636475,
      "learning_rate": 5.725845599030438e-05,
      "loss": 1.2953,
      "mean_token_accuracy": 0.754032127559185,
      "num_tokens": 1902502399.0,
      "step": 26470
    },
    {
      "entropy": 0.8696189224720001,
      "epoch": 7.217225402016898,
      "grad_norm": 0.1460484117269516,
      "learning_rate": 5.722872883097009e-05,
      "loss": 1.2962,
      "mean_token_accuracy": 0.7574547484517098,
      "num_tokens": 1903209640.0,
      "step": 26480
    },
    {
      "entropy": 0.8711689874529839,
      "epoch": 7.219950940310711,
      "grad_norm": 0.1420704871416092,
      "learning_rate": 5.719899906163485e-05,
      "loss": 1.2968,
      "mean_token_accuracy": 0.7579603239893913,
      "num_tokens": 1903931633.0,
      "step": 26490
    },
    {
      "entropy": 0.8681849867105484,
      "epoch": 7.222676478604525,
      "grad_norm": 0.1611538976430893,
      "learning_rate": 5.716926669303285e-05,
      "loss": 1.3001,
      "mean_token_accuracy": 0.7574627697467804,
      "num_tokens": 1904635178.0,
      "step": 26500
    },
    {
      "entropy": 0.8808808535337448,
      "epoch": 7.225402016898338,
      "grad_norm": 0.14451847970485687,
      "learning_rate": 5.713953173589925e-05,
      "loss": 1.3024,
      "mean_token_accuracy": 0.7549888685345649,
      "num_tokens": 1905359027.0,
      "step": 26510
    },
    {
      "entropy": 0.880359698832035,
      "epoch": 7.228127555192151,
      "grad_norm": 0.1502303034067154,
      "learning_rate": 5.710979420097018e-05,
      "loss": 1.3079,
      "mean_token_accuracy": 0.7542410060763359,
      "num_tokens": 1906070571.0,
      "step": 26520
    },
    {
      "entropy": 0.8744180366396904,
      "epoch": 7.230853093485964,
      "grad_norm": 0.1395016610622406,
      "learning_rate": 5.708005409898259e-05,
      "loss": 1.2944,
      "mean_token_accuracy": 0.7585380360484123,
      "num_tokens": 1906801705.0,
      "step": 26530
    },
    {
      "entropy": 0.8749017134308815,
      "epoch": 7.233578631779777,
      "grad_norm": 0.15864379703998566,
      "learning_rate": 5.7050311440674476e-05,
      "loss": 1.3022,
      "mean_token_accuracy": 0.7603865161538124,
      "num_tokens": 1907533354.0,
      "step": 26540
    },
    {
      "entropy": 0.872776010632515,
      "epoch": 7.2363041700735895,
      "grad_norm": 0.151897132396698,
      "learning_rate": 5.702056623678469e-05,
      "loss": 1.2926,
      "mean_token_accuracy": 0.7577180013060569,
      "num_tokens": 1908260593.0,
      "step": 26550
    },
    {
      "entropy": 0.8791093736886978,
      "epoch": 7.2390297083674024,
      "grad_norm": 0.14768271148204803,
      "learning_rate": 5.6990818498053014e-05,
      "loss": 1.3086,
      "mean_token_accuracy": 0.753087967634201,
      "num_tokens": 1908978048.0,
      "step": 26560
    },
    {
      "entropy": 0.8763909667730332,
      "epoch": 7.241755246661215,
      "grad_norm": 0.1572977751493454,
      "learning_rate": 5.696106823522016e-05,
      "loss": 1.2934,
      "mean_token_accuracy": 0.755253580212593,
      "num_tokens": 1909697715.0,
      "step": 26570
    },
    {
      "entropy": 0.8799683541059494,
      "epoch": 7.244480784955028,
      "grad_norm": 0.14823664724826813,
      "learning_rate": 5.693131545902774e-05,
      "loss": 1.3091,
      "mean_token_accuracy": 0.7538741514086723,
      "num_tokens": 1910411537.0,
      "step": 26580
    },
    {
      "entropy": 0.8791606545448303,
      "epoch": 7.247206323248841,
      "grad_norm": 0.1453849971294403,
      "learning_rate": 5.690156018021827e-05,
      "loss": 1.3042,
      "mean_token_accuracy": 0.7514085024595261,
      "num_tokens": 1911122559.0,
      "step": 26590
    },
    {
      "entropy": 0.8769569367170333,
      "epoch": 7.249931861542654,
      "grad_norm": 0.1487053632736206,
      "learning_rate": 5.687180240953518e-05,
      "loss": 1.3013,
      "mean_token_accuracy": 0.7544289767742157,
      "num_tokens": 1911838374.0,
      "step": 26600
    },
    {
      "entropy": 0.8753040879964828,
      "epoch": 7.252657399836468,
      "grad_norm": 0.1425638496875763,
      "learning_rate": 5.684204215772281e-05,
      "loss": 1.3008,
      "mean_token_accuracy": 0.757389535009861,
      "num_tokens": 1912564686.0,
      "step": 26610
    },
    {
      "entropy": 0.8770721167325973,
      "epoch": 7.255382938130281,
      "grad_norm": 0.14608031511306763,
      "learning_rate": 5.681227943552634e-05,
      "loss": 1.3003,
      "mean_token_accuracy": 0.75359016507864,
      "num_tokens": 1913285484.0,
      "step": 26620
    },
    {
      "entropy": 0.8745535492897034,
      "epoch": 7.258108476424094,
      "grad_norm": 0.14657150208950043,
      "learning_rate": 5.678251425369191e-05,
      "loss": 1.2932,
      "mean_token_accuracy": 0.7575717538595199,
      "num_tokens": 1914006490.0,
      "step": 26630
    },
    {
      "entropy": 0.8714345395565033,
      "epoch": 7.260834014717907,
      "grad_norm": 0.1463703066110611,
      "learning_rate": 5.675274662296654e-05,
      "loss": 1.2988,
      "mean_token_accuracy": 0.7547225907444954,
      "num_tokens": 1914717634.0,
      "step": 26640
    },
    {
      "entropy": 0.8697472095489502,
      "epoch": 7.26355955301172,
      "grad_norm": 0.1437256783246994,
      "learning_rate": 5.672297655409806e-05,
      "loss": 1.293,
      "mean_token_accuracy": 0.7568093731999397,
      "num_tokens": 1915443995.0,
      "step": 26650
    },
    {
      "entropy": 0.8644485801458359,
      "epoch": 7.266285091305533,
      "grad_norm": 0.14560437202453613,
      "learning_rate": 5.669320405783528e-05,
      "loss": 1.2904,
      "mean_token_accuracy": 0.7605119436979294,
      "num_tokens": 1916162662.0,
      "step": 26660
    },
    {
      "entropy": 0.8658681184053421,
      "epoch": 7.269010629599346,
      "grad_norm": 0.15063898265361786,
      "learning_rate": 5.666342914492785e-05,
      "loss": 1.2954,
      "mean_token_accuracy": 0.7597482159733773,
      "num_tokens": 1916885796.0,
      "step": 26670
    },
    {
      "entropy": 0.8673249945044518,
      "epoch": 7.271736167893159,
      "grad_norm": 0.14441640675067902,
      "learning_rate": 5.6633651826126235e-05,
      "loss": 1.2921,
      "mean_token_accuracy": 0.759340013563633,
      "num_tokens": 1917602591.0,
      "step": 26680
    },
    {
      "entropy": 0.8810052171349525,
      "epoch": 7.274461706186972,
      "grad_norm": 0.1468001902103424,
      "learning_rate": 5.6603872112181855e-05,
      "loss": 1.3001,
      "mean_token_accuracy": 0.7559056729078293,
      "num_tokens": 1918324243.0,
      "step": 26690
    },
    {
      "entropy": 0.873295296728611,
      "epoch": 7.277187244480785,
      "grad_norm": 0.1526554822921753,
      "learning_rate": 5.6574090013846946e-05,
      "loss": 1.29,
      "mean_token_accuracy": 0.7562502101063728,
      "num_tokens": 1919042536.0,
      "step": 26700
    },
    {
      "entropy": 0.8708538606762886,
      "epoch": 7.279912782774598,
      "grad_norm": 0.16268682479858398,
      "learning_rate": 5.654430554187462e-05,
      "loss": 1.297,
      "mean_token_accuracy": 0.7591925546526909,
      "num_tokens": 1919774068.0,
      "step": 26710
    },
    {
      "entropy": 0.8741784408688545,
      "epoch": 7.282638321068411,
      "grad_norm": 0.13987214863300323,
      "learning_rate": 5.651451870701884e-05,
      "loss": 1.2991,
      "mean_token_accuracy": 0.7585245609283447,
      "num_tokens": 1920497162.0,
      "step": 26720
    },
    {
      "entropy": 0.8752786368131638,
      "epoch": 7.285363859362224,
      "grad_norm": 0.14615780115127563,
      "learning_rate": 5.648472952003443e-05,
      "loss": 1.3053,
      "mean_token_accuracy": 0.7574028953909874,
      "num_tokens": 1921217946.0,
      "step": 26730
    },
    {
      "entropy": 0.872516205906868,
      "epoch": 7.288089397656037,
      "grad_norm": 0.1476518213748932,
      "learning_rate": 5.6454937991677045e-05,
      "loss": 1.295,
      "mean_token_accuracy": 0.7549552768468857,
      "num_tokens": 1921927093.0,
      "step": 26740
    },
    {
      "entropy": 0.8677942544221878,
      "epoch": 7.29081493594985,
      "grad_norm": 0.14972105622291565,
      "learning_rate": 5.642514413270321e-05,
      "loss": 1.2867,
      "mean_token_accuracy": 0.7599112004041672,
      "num_tokens": 1922658036.0,
      "step": 26750
    },
    {
      "entropy": 0.8700306206941605,
      "epoch": 7.293540474243663,
      "grad_norm": 0.15284699201583862,
      "learning_rate": 5.639534795387029e-05,
      "loss": 1.2971,
      "mean_token_accuracy": 0.7596366837620735,
      "num_tokens": 1923385377.0,
      "step": 26760
    },
    {
      "entropy": 0.8781607255339623,
      "epoch": 7.296266012537476,
      "grad_norm": 0.14511984586715698,
      "learning_rate": 5.6365549465936455e-05,
      "loss": 1.2881,
      "mean_token_accuracy": 0.7567567452788353,
      "num_tokens": 1924115821.0,
      "step": 26770
    },
    {
      "entropy": 0.8762117221951484,
      "epoch": 7.298991550831289,
      "grad_norm": 0.1574789583683014,
      "learning_rate": 5.633574867966073e-05,
      "loss": 1.298,
      "mean_token_accuracy": 0.7559019878506661,
      "num_tokens": 1924834287.0,
      "step": 26780
    },
    {
      "entropy": 0.8820765510201454,
      "epoch": 7.301717089125102,
      "grad_norm": 0.14718078076839447,
      "learning_rate": 5.630594560580298e-05,
      "loss": 1.2951,
      "mean_token_accuracy": 0.7587728559970855,
      "num_tokens": 1925575763.0,
      "step": 26790
    },
    {
      "entropy": 0.8790650695562363,
      "epoch": 7.304442627418915,
      "grad_norm": 0.15892545878887177,
      "learning_rate": 5.62761402551239e-05,
      "loss": 1.2928,
      "mean_token_accuracy": 0.7535487607121467,
      "num_tokens": 1926296917.0,
      "step": 26800
    },
    {
      "entropy": 0.8677218273282051,
      "epoch": 7.307168165712728,
      "grad_norm": 0.13702869415283203,
      "learning_rate": 5.6246332638384966e-05,
      "loss": 1.2863,
      "mean_token_accuracy": 0.7585681483149529,
      "num_tokens": 1927027864.0,
      "step": 26810
    },
    {
      "entropy": 0.8745145723223686,
      "epoch": 7.309893704006542,
      "grad_norm": 0.1458585262298584,
      "learning_rate": 5.6216522766348546e-05,
      "loss": 1.3009,
      "mean_token_accuracy": 0.7565836235880852,
      "num_tokens": 1927750164.0,
      "step": 26820
    },
    {
      "entropy": 0.8716302827000618,
      "epoch": 7.312619242300355,
      "grad_norm": 0.14542189240455627,
      "learning_rate": 5.618671064977772e-05,
      "loss": 1.2965,
      "mean_token_accuracy": 0.7582589849829674,
      "num_tokens": 1928472162.0,
      "step": 26830
    },
    {
      "entropy": 0.8728285521268845,
      "epoch": 7.315344780594168,
      "grad_norm": 0.1470698118209839,
      "learning_rate": 5.615689629943647e-05,
      "loss": 1.2997,
      "mean_token_accuracy": 0.7578873261809349,
      "num_tokens": 1929196879.0,
      "step": 26840
    },
    {
      "entropy": 0.8646959900856018,
      "epoch": 7.3180703188879805,
      "grad_norm": 0.14543084800243378,
      "learning_rate": 5.612707972608955e-05,
      "loss": 1.296,
      "mean_token_accuracy": 0.7581882014870643,
      "num_tokens": 1929906321.0,
      "step": 26850
    },
    {
      "entropy": 0.8686492219567299,
      "epoch": 7.3207958571817935,
      "grad_norm": 0.15842905640602112,
      "learning_rate": 5.60972609405025e-05,
      "loss": 1.2863,
      "mean_token_accuracy": 0.7616284027695656,
      "num_tokens": 1930641616.0,
      "step": 26860
    },
    {
      "entropy": 0.8771625027060509,
      "epoch": 7.323521395475606,
      "grad_norm": 0.14808574318885803,
      "learning_rate": 5.606743995344167e-05,
      "loss": 1.2988,
      "mean_token_accuracy": 0.7534963071346283,
      "num_tokens": 1931348000.0,
      "step": 26870
    },
    {
      "entropy": 0.8703848287463188,
      "epoch": 7.326246933769419,
      "grad_norm": 0.14173507690429688,
      "learning_rate": 5.603761677567424e-05,
      "loss": 1.2915,
      "mean_token_accuracy": 0.7567645877599716,
      "num_tokens": 1932061448.0,
      "step": 26880
    },
    {
      "entropy": 0.875830689072609,
      "epoch": 7.328972472063232,
      "grad_norm": 0.14382077753543854,
      "learning_rate": 5.6007791417968114e-05,
      "loss": 1.3019,
      "mean_token_accuracy": 0.7579288348555565,
      "num_tokens": 1932790003.0,
      "step": 26890
    },
    {
      "entropy": 0.8764667242765427,
      "epoch": 7.331698010357045,
      "grad_norm": 0.1480296105146408,
      "learning_rate": 5.597796389109205e-05,
      "loss": 1.305,
      "mean_token_accuracy": 0.7592344105243682,
      "num_tokens": 1933516726.0,
      "step": 26900
    },
    {
      "entropy": 0.8774703487753868,
      "epoch": 7.334423548650858,
      "grad_norm": 0.15020212531089783,
      "learning_rate": 5.594813420581554e-05,
      "loss": 1.3004,
      "mean_token_accuracy": 0.7581730052828789,
      "num_tokens": 1934247798.0,
      "step": 26910
    },
    {
      "entropy": 0.8736540108919144,
      "epoch": 7.337149086944671,
      "grad_norm": 0.15321451425552368,
      "learning_rate": 5.591830237290887e-05,
      "loss": 1.2876,
      "mean_token_accuracy": 0.7563091889023781,
      "num_tokens": 1934980623.0,
      "step": 26920
    },
    {
      "entropy": 0.8708978682756424,
      "epoch": 7.339874625238485,
      "grad_norm": 0.14175894856452942,
      "learning_rate": 5.588846840314313e-05,
      "loss": 1.2889,
      "mean_token_accuracy": 0.7525080293416977,
      "num_tokens": 1935691312.0,
      "step": 26930
    },
    {
      "entropy": 0.8693087965250015,
      "epoch": 7.342600163532298,
      "grad_norm": 0.1484062820672989,
      "learning_rate": 5.585863230729015e-05,
      "loss": 1.2924,
      "mean_token_accuracy": 0.7605011075735092,
      "num_tokens": 1936412220.0,
      "step": 26940
    },
    {
      "entropy": 0.8658810436725617,
      "epoch": 7.345325701826111,
      "grad_norm": 0.14559951424598694,
      "learning_rate": 5.582879409612252e-05,
      "loss": 1.2841,
      "mean_token_accuracy": 0.7626533910632134,
      "num_tokens": 1937137010.0,
      "step": 26950
    },
    {
      "entropy": 0.8778633654117585,
      "epoch": 7.348051240119924,
      "grad_norm": 0.13812687993049622,
      "learning_rate": 5.57989537804136e-05,
      "loss": 1.2947,
      "mean_token_accuracy": 0.7557541623711586,
      "num_tokens": 1937853165.0,
      "step": 26960
    },
    {
      "entropy": 0.8690854504704475,
      "epoch": 7.350776778413737,
      "grad_norm": 0.1555824875831604,
      "learning_rate": 5.576911137093755e-05,
      "loss": 1.2918,
      "mean_token_accuracy": 0.7585563018918038,
      "num_tokens": 1938584116.0,
      "step": 26970
    },
    {
      "entropy": 0.8780083298683167,
      "epoch": 7.35350231670755,
      "grad_norm": 0.1554984301328659,
      "learning_rate": 5.5739266878469244e-05,
      "loss": 1.2906,
      "mean_token_accuracy": 0.7591735869646072,
      "num_tokens": 1939314489.0,
      "step": 26980
    },
    {
      "entropy": 0.8686507850885391,
      "epoch": 7.356227855001363,
      "grad_norm": 0.1445007026195526,
      "learning_rate": 5.570942031378431e-05,
      "loss": 1.2948,
      "mean_token_accuracy": 0.7603983774781227,
      "num_tokens": 1940039701.0,
      "step": 26990
    },
    {
      "entropy": 0.8691764920949936,
      "epoch": 7.358953393295176,
      "grad_norm": 0.1479530930519104,
      "learning_rate": 5.567957168765914e-05,
      "loss": 1.2934,
      "mean_token_accuracy": 0.757313358783722,
      "num_tokens": 1940745127.0,
      "step": 27000
    },
    {
      "entropy": 0.8695494890213012,
      "epoch": 7.361678931588989,
      "grad_norm": 0.16647686064243317,
      "learning_rate": 5.564972101087085e-05,
      "loss": 1.2994,
      "mean_token_accuracy": 0.7575660184025764,
      "num_tokens": 1941473340.0,
      "step": 27010
    },
    {
      "entropy": 0.8801256656646729,
      "epoch": 7.3644044698828015,
      "grad_norm": 0.16509966552257538,
      "learning_rate": 5.561986829419731e-05,
      "loss": 1.299,
      "mean_token_accuracy": 0.7550579234957695,
      "num_tokens": 1942196825.0,
      "step": 27020
    },
    {
      "entropy": 0.87327411621809,
      "epoch": 7.3671300081766145,
      "grad_norm": 0.1618693321943283,
      "learning_rate": 5.5590013548417144e-05,
      "loss": 1.3043,
      "mean_token_accuracy": 0.7550663352012634,
      "num_tokens": 1942911042.0,
      "step": 27030
    },
    {
      "entropy": 0.8774277091026306,
      "epoch": 7.369855546470428,
      "grad_norm": 0.20013737678527832,
      "learning_rate": 5.556015678430968e-05,
      "loss": 1.2973,
      "mean_token_accuracy": 0.7544228315353394,
      "num_tokens": 1943632151.0,
      "step": 27040
    },
    {
      "entropy": 0.8747197523713112,
      "epoch": 7.372581084764241,
      "grad_norm": 0.1496356576681137,
      "learning_rate": 5.553029801265499e-05,
      "loss": 1.3001,
      "mean_token_accuracy": 0.757723331451416,
      "num_tokens": 1944350133.0,
      "step": 27050
    },
    {
      "entropy": 0.8781649515032768,
      "epoch": 7.375306623058054,
      "grad_norm": 0.14166666567325592,
      "learning_rate": 5.550043724423386e-05,
      "loss": 1.2967,
      "mean_token_accuracy": 0.7560077399015427,
      "num_tokens": 1945065543.0,
      "step": 27060
    },
    {
      "entropy": 0.8705385491251946,
      "epoch": 7.378032161351867,
      "grad_norm": 0.14302848279476166,
      "learning_rate": 5.547057448982781e-05,
      "loss": 1.299,
      "mean_token_accuracy": 0.7566994354128838,
      "num_tokens": 1945786897.0,
      "step": 27070
    },
    {
      "entropy": 0.8793146178126335,
      "epoch": 7.38075769964568,
      "grad_norm": 0.15471141040325165,
      "learning_rate": 5.544070976021907e-05,
      "loss": 1.2983,
      "mean_token_accuracy": 0.7566537842154503,
      "num_tokens": 1946507446.0,
      "step": 27080
    },
    {
      "entropy": 0.8735587447881699,
      "epoch": 7.383483237939493,
      "grad_norm": 0.1450737714767456,
      "learning_rate": 5.541084306619056e-05,
      "loss": 1.3035,
      "mean_token_accuracy": 0.7568038001656532,
      "num_tokens": 1947216149.0,
      "step": 27090
    },
    {
      "entropy": 0.8783910438418389,
      "epoch": 7.386208776233306,
      "grad_norm": 0.14964261651039124,
      "learning_rate": 5.5380974418525965e-05,
      "loss": 1.3089,
      "mean_token_accuracy": 0.7551267296075821,
      "num_tokens": 1947932134.0,
      "step": 27100
    },
    {
      "entropy": 0.8729043811559677,
      "epoch": 7.388934314527119,
      "grad_norm": 0.15176354348659515,
      "learning_rate": 5.535110382800963e-05,
      "loss": 1.3021,
      "mean_token_accuracy": 0.7548694238066673,
      "num_tokens": 1948644380.0,
      "step": 27110
    },
    {
      "entropy": 0.8665193617343903,
      "epoch": 7.391659852820932,
      "grad_norm": 0.15515904128551483,
      "learning_rate": 5.5321231305426616e-05,
      "loss": 1.2964,
      "mean_token_accuracy": 0.7560141786932946,
      "num_tokens": 1949344586.0,
      "step": 27120
    },
    {
      "entropy": 0.8758058831095695,
      "epoch": 7.394385391114745,
      "grad_norm": 0.14520084857940674,
      "learning_rate": 5.529135686156266e-05,
      "loss": 1.307,
      "mean_token_accuracy": 0.7543540596961975,
      "num_tokens": 1950046275.0,
      "step": 27130
    },
    {
      "entropy": 0.8598334476351738,
      "epoch": 7.397110929408559,
      "grad_norm": 0.14953961968421936,
      "learning_rate": 5.5261480507204255e-05,
      "loss": 1.2895,
      "mean_token_accuracy": 0.7635306343436241,
      "num_tokens": 1950777609.0,
      "step": 27140
    },
    {
      "entropy": 0.8784276217222213,
      "epoch": 7.399836467702372,
      "grad_norm": 0.14264141023159027,
      "learning_rate": 5.52316022531385e-05,
      "loss": 1.3017,
      "mean_token_accuracy": 0.756753858923912,
      "num_tokens": 1951504210.0,
      "step": 27150
    },
    {
      "entropy": 0.8731119826436042,
      "epoch": 7.4025620059961845,
      "grad_norm": 0.14532233774662018,
      "learning_rate": 5.5201722110153244e-05,
      "loss": 1.2976,
      "mean_token_accuracy": 0.7603534251451493,
      "num_tokens": 1952236971.0,
      "step": 27160
    },
    {
      "entropy": 0.8778833121061325,
      "epoch": 7.4052875442899975,
      "grad_norm": 0.15055438876152039,
      "learning_rate": 5.517184008903699e-05,
      "loss": 1.2967,
      "mean_token_accuracy": 0.7536444157361984,
      "num_tokens": 1952944489.0,
      "step": 27170
    },
    {
      "entropy": 0.88260188549757,
      "epoch": 7.40801308258381,
      "grad_norm": 0.14573059976100922,
      "learning_rate": 5.5141956200578914e-05,
      "loss": 1.3108,
      "mean_token_accuracy": 0.7570423111319542,
      "num_tokens": 1953671895.0,
      "step": 27180
    },
    {
      "entropy": 0.871610289812088,
      "epoch": 7.410738620877623,
      "grad_norm": 0.14601604640483856,
      "learning_rate": 5.51120704555689e-05,
      "loss": 1.2873,
      "mean_token_accuracy": 0.758485485613346,
      "num_tokens": 1954401100.0,
      "step": 27190
    },
    {
      "entropy": 0.8706778183579444,
      "epoch": 7.413464159171436,
      "grad_norm": 0.17026714980602264,
      "learning_rate": 5.508218286479746e-05,
      "loss": 1.2944,
      "mean_token_accuracy": 0.7586376383900643,
      "num_tokens": 1955119071.0,
      "step": 27200
    },
    {
      "entropy": 0.8755591064691544,
      "epoch": 7.416189697465249,
      "grad_norm": 0.13916918635368347,
      "learning_rate": 5.5052293439055804e-05,
      "loss": 1.2944,
      "mean_token_accuracy": 0.7645180627703667,
      "num_tokens": 1955867848.0,
      "step": 27210
    },
    {
      "entropy": 0.8669392347335816,
      "epoch": 7.418915235759062,
      "grad_norm": 0.21698226034641266,
      "learning_rate": 5.502240218913577e-05,
      "loss": 1.2882,
      "mean_token_accuracy": 0.7592273488640785,
      "num_tokens": 1956583663.0,
      "step": 27220
    },
    {
      "entropy": 0.870184437930584,
      "epoch": 7.421640774052875,
      "grad_norm": 0.14668874442577362,
      "learning_rate": 5.499250912582992e-05,
      "loss": 1.2976,
      "mean_token_accuracy": 0.75679152905941,
      "num_tokens": 1957298917.0,
      "step": 27230
    },
    {
      "entropy": 0.8762162327766418,
      "epoch": 7.424366312346688,
      "grad_norm": 0.150191068649292,
      "learning_rate": 5.4962614259931365e-05,
      "loss": 1.3014,
      "mean_token_accuracy": 0.7538296952843666,
      "num_tokens": 1958019543.0,
      "step": 27240
    },
    {
      "entropy": 0.8609578981995583,
      "epoch": 7.427091850640501,
      "grad_norm": 0.15371190011501312,
      "learning_rate": 5.4932717602233985e-05,
      "loss": 1.2919,
      "mean_token_accuracy": 0.7573593899607658,
      "num_tokens": 1958725145.0,
      "step": 27250
    },
    {
      "entropy": 0.8777864828705788,
      "epoch": 7.429817388934315,
      "grad_norm": 0.1603131741285324,
      "learning_rate": 5.4902819163532224e-05,
      "loss": 1.3016,
      "mean_token_accuracy": 0.7558649733662606,
      "num_tokens": 1959453162.0,
      "step": 27260
    },
    {
      "entropy": 0.8782357349991798,
      "epoch": 7.432542927228128,
      "grad_norm": 0.14163288474082947,
      "learning_rate": 5.48729189546212e-05,
      "loss": 1.3043,
      "mean_token_accuracy": 0.7551824197173118,
      "num_tokens": 1960167109.0,
      "step": 27270
    },
    {
      "entropy": 0.8796595558524132,
      "epoch": 7.435268465521941,
      "grad_norm": 0.1622946560382843,
      "learning_rate": 5.484301698629667e-05,
      "loss": 1.3034,
      "mean_token_accuracy": 0.7564584419131279,
      "num_tokens": 1960888368.0,
      "step": 27280
    },
    {
      "entropy": 0.8698703140020371,
      "epoch": 7.437994003815754,
      "grad_norm": 0.14658711850643158,
      "learning_rate": 5.4813113269355035e-05,
      "loss": 1.3043,
      "mean_token_accuracy": 0.7558253839612007,
      "num_tokens": 1961592155.0,
      "step": 27290
    },
    {
      "entropy": 0.8798405855894089,
      "epoch": 7.440719542109567,
      "grad_norm": 0.1534138321876526,
      "learning_rate": 5.4783207814593286e-05,
      "loss": 1.3005,
      "mean_token_accuracy": 0.7533506646752357,
      "num_tokens": 1962301630.0,
      "step": 27300
    },
    {
      "entropy": 0.8711357161402702,
      "epoch": 7.44344508040338,
      "grad_norm": 0.1563558578491211,
      "learning_rate": 5.475330063280909e-05,
      "loss": 1.2939,
      "mean_token_accuracy": 0.7589615762233735,
      "num_tokens": 1963029478.0,
      "step": 27310
    },
    {
      "entropy": 0.874621294438839,
      "epoch": 7.4461706186971925,
      "grad_norm": 0.17626208066940308,
      "learning_rate": 5.472339173480072e-05,
      "loss": 1.2976,
      "mean_token_accuracy": 0.762251828610897,
      "num_tokens": 1963755758.0,
      "step": 27320
    },
    {
      "entropy": 0.8707356587052345,
      "epoch": 7.4488961569910055,
      "grad_norm": 0.14620471000671387,
      "learning_rate": 5.4693481131367053e-05,
      "loss": 1.2902,
      "mean_token_accuracy": 0.760369473695755,
      "num_tokens": 1964482144.0,
      "step": 27330
    },
    {
      "entropy": 0.8790704697370529,
      "epoch": 7.451621695284818,
      "grad_norm": 0.1517985761165619,
      "learning_rate": 5.466356883330761e-05,
      "loss": 1.3049,
      "mean_token_accuracy": 0.7572114810347557,
      "num_tokens": 1965187783.0,
      "step": 27340
    },
    {
      "entropy": 0.8741227284073829,
      "epoch": 7.454347233578631,
      "grad_norm": 0.15087679028511047,
      "learning_rate": 5.46336548514225e-05,
      "loss": 1.3024,
      "mean_token_accuracy": 0.7526372775435448,
      "num_tokens": 1965896017.0,
      "step": 27350
    },
    {
      "entropy": 0.8820385470986366,
      "epoch": 7.457072771872445,
      "grad_norm": 0.1478431671857834,
      "learning_rate": 5.460373919651245e-05,
      "loss": 1.304,
      "mean_token_accuracy": 0.7544811740517616,
      "num_tokens": 1966614752.0,
      "step": 27360
    },
    {
      "entropy": 0.86817337423563,
      "epoch": 7.459798310166258,
      "grad_norm": 0.14179888367652893,
      "learning_rate": 5.45738218793788e-05,
      "loss": 1.2929,
      "mean_token_accuracy": 0.7584007635712624,
      "num_tokens": 1967326882.0,
      "step": 27370
    },
    {
      "entropy": 0.8750143766403198,
      "epoch": 7.462523848460071,
      "grad_norm": 0.15783092379570007,
      "learning_rate": 5.454390291082346e-05,
      "loss": 1.306,
      "mean_token_accuracy": 0.7565370664000511,
      "num_tokens": 1968045004.0,
      "step": 27380
    },
    {
      "entropy": 0.8752815082669259,
      "epoch": 7.465249386753884,
      "grad_norm": 0.143658846616745,
      "learning_rate": 5.451398230164897e-05,
      "loss": 1.2927,
      "mean_token_accuracy": 0.7576921477913856,
      "num_tokens": 1968757703.0,
      "step": 27390
    },
    {
      "entropy": 0.8686357542872429,
      "epoch": 7.467974925047697,
      "grad_norm": 0.16239280998706818,
      "learning_rate": 5.448406006265846e-05,
      "loss": 1.2898,
      "mean_token_accuracy": 0.7560498282313347,
      "num_tokens": 1969467259.0,
      "step": 27400
    },
    {
      "entropy": 0.8754814326763153,
      "epoch": 7.47070046334151,
      "grad_norm": 0.14713862538337708,
      "learning_rate": 5.4454136204655624e-05,
      "loss": 1.2993,
      "mean_token_accuracy": 0.7532849207520484,
      "num_tokens": 1970171090.0,
      "step": 27410
    },
    {
      "entropy": 0.878950223326683,
      "epoch": 7.473426001635323,
      "grad_norm": 0.14513806998729706,
      "learning_rate": 5.442421073844475e-05,
      "loss": 1.2991,
      "mean_token_accuracy": 0.7543308869004249,
      "num_tokens": 1970875920.0,
      "step": 27420
    },
    {
      "entropy": 0.8695757389068604,
      "epoch": 7.476151539929136,
      "grad_norm": 0.1472143828868866,
      "learning_rate": 5.439428367483072e-05,
      "loss": 1.281,
      "mean_token_accuracy": 0.75978524684906,
      "num_tokens": 1971613157.0,
      "step": 27430
    },
    {
      "entropy": 0.8697513654828072,
      "epoch": 7.478877078222949,
      "grad_norm": 0.14886444807052612,
      "learning_rate": 5.436435502461897e-05,
      "loss": 1.2883,
      "mean_token_accuracy": 0.7608008369803428,
      "num_tokens": 1972352268.0,
      "step": 27440
    },
    {
      "entropy": 0.8863059863448143,
      "epoch": 7.481602616516762,
      "grad_norm": 0.15480509400367737,
      "learning_rate": 5.4334424798615546e-05,
      "loss": 1.3142,
      "mean_token_accuracy": 0.7498865917325019,
      "num_tokens": 1973057486.0,
      "step": 27450
    },
    {
      "entropy": 0.870640966296196,
      "epoch": 7.484328154810575,
      "grad_norm": 0.1481386125087738,
      "learning_rate": 5.430449300762702e-05,
      "loss": 1.3039,
      "mean_token_accuracy": 0.7583491995930671,
      "num_tokens": 1973770465.0,
      "step": 27460
    },
    {
      "entropy": 0.8789900720119477,
      "epoch": 7.4870536931043885,
      "grad_norm": 0.14939655363559723,
      "learning_rate": 5.427455966246057e-05,
      "loss": 1.3076,
      "mean_token_accuracy": 0.7516216769814491,
      "num_tokens": 1974475301.0,
      "step": 27470
    },
    {
      "entropy": 0.8683831721544266,
      "epoch": 7.4897792313982015,
      "grad_norm": 0.14119526743888855,
      "learning_rate": 5.424462477392388e-05,
      "loss": 1.292,
      "mean_token_accuracy": 0.7570024654269218,
      "num_tokens": 1975181148.0,
      "step": 27480
    },
    {
      "entropy": 0.8829577803611756,
      "epoch": 7.492504769692014,
      "grad_norm": 0.15738560259342194,
      "learning_rate": 5.421468835282524e-05,
      "loss": 1.3134,
      "mean_token_accuracy": 0.751892414689064,
      "num_tokens": 1975894738.0,
      "step": 27490
    },
    {
      "entropy": 0.8754595801234245,
      "epoch": 7.495230307985827,
      "grad_norm": 0.16415882110595703,
      "learning_rate": 5.418475040997348e-05,
      "loss": 1.3025,
      "mean_token_accuracy": 0.7552436009049416,
      "num_tokens": 1976604250.0,
      "step": 27500
    },
    {
      "entropy": 0.8759715840220451,
      "epoch": 7.49795584627964,
      "grad_norm": 0.15219652652740479,
      "learning_rate": 5.4154810956177994e-05,
      "loss": 1.3003,
      "mean_token_accuracy": 0.7568041488528252,
      "num_tokens": 1977319582.0,
      "step": 27510
    },
    {
      "entropy": 0.874935844540596,
      "epoch": 7.500681384573453,
      "grad_norm": 0.1453443467617035,
      "learning_rate": 5.412487000224868e-05,
      "loss": 1.2957,
      "mean_token_accuracy": 0.7554998025298119,
      "num_tokens": 1978036600.0,
      "step": 27520
    },
    {
      "entropy": 0.8738935187458992,
      "epoch": 7.503406922867266,
      "grad_norm": 0.146016925573349,
      "learning_rate": 5.409492755899602e-05,
      "loss": 1.2973,
      "mean_token_accuracy": 0.7545162200927734,
      "num_tokens": 1978750720.0,
      "step": 27530
    },
    {
      "entropy": 0.8722354531288147,
      "epoch": 7.506132461161079,
      "grad_norm": 0.14861078560352325,
      "learning_rate": 5.4064983637231016e-05,
      "loss": 1.3033,
      "mean_token_accuracy": 0.7554042488336563,
      "num_tokens": 1979460934.0,
      "step": 27540
    },
    {
      "entropy": 0.8751856788992882,
      "epoch": 7.508857999454892,
      "grad_norm": 0.15488703548908234,
      "learning_rate": 5.403503824776518e-05,
      "loss": 1.2938,
      "mean_token_accuracy": 0.7567870914936066,
      "num_tokens": 1980177931.0,
      "step": 27550
    },
    {
      "entropy": 0.8693725138902664,
      "epoch": 7.511583537748705,
      "grad_norm": 0.17622137069702148,
      "learning_rate": 5.400509140141061e-05,
      "loss": 1.2986,
      "mean_token_accuracy": 0.7618825286626816,
      "num_tokens": 1980909112.0,
      "step": 27560
    },
    {
      "entropy": 0.875079782307148,
      "epoch": 7.514309076042519,
      "grad_norm": 0.14566169679164886,
      "learning_rate": 5.397514310897988e-05,
      "loss": 1.2971,
      "mean_token_accuracy": 0.7558818951249122,
      "num_tokens": 1981630452.0,
      "step": 27570
    },
    {
      "entropy": 0.8672457680106163,
      "epoch": 7.517034614336332,
      "grad_norm": 0.15224742889404297,
      "learning_rate": 5.3945193381286095e-05,
      "loss": 1.2867,
      "mean_token_accuracy": 0.7568916633725167,
      "num_tokens": 1982352311.0,
      "step": 27580
    },
    {
      "entropy": 0.8709974974393845,
      "epoch": 7.519760152630145,
      "grad_norm": 0.14790773391723633,
      "learning_rate": 5.391524222914293e-05,
      "loss": 1.299,
      "mean_token_accuracy": 0.7587826490402222,
      "num_tokens": 1983071543.0,
      "step": 27590
    },
    {
      "entropy": 0.8683429807424545,
      "epoch": 7.522485690923958,
      "grad_norm": 0.14690671861171722,
      "learning_rate": 5.388528966336448e-05,
      "loss": 1.3012,
      "mean_token_accuracy": 0.7594067826867104,
      "num_tokens": 1983786649.0,
      "step": 27600
    },
    {
      "entropy": 0.8661997959017753,
      "epoch": 7.525211229217771,
      "grad_norm": 0.16435834765434265,
      "learning_rate": 5.385533569476543e-05,
      "loss": 1.2879,
      "mean_token_accuracy": 0.7576445356011391,
      "num_tokens": 1984501021.0,
      "step": 27610
    },
    {
      "entropy": 0.8720558568835258,
      "epoch": 7.527936767511584,
      "grad_norm": 0.15992335975170135,
      "learning_rate": 5.3825380334160955e-05,
      "loss": 1.293,
      "mean_token_accuracy": 0.7589874520897866,
      "num_tokens": 1985223828.0,
      "step": 27620
    },
    {
      "entropy": 0.8726562917232513,
      "epoch": 7.5306623058053965,
      "grad_norm": 0.15062344074249268,
      "learning_rate": 5.379542359236669e-05,
      "loss": 1.2995,
      "mean_token_accuracy": 0.7584767267107964,
      "num_tokens": 1985941467.0,
      "step": 27630
    },
    {
      "entropy": 0.8708229914307595,
      "epoch": 7.5333878440992095,
      "grad_norm": 0.15603512525558472,
      "learning_rate": 5.376546548019882e-05,
      "loss": 1.2933,
      "mean_token_accuracy": 0.7633575603365899,
      "num_tokens": 1986676136.0,
      "step": 27640
    },
    {
      "entropy": 0.8802310168743134,
      "epoch": 7.536113382393022,
      "grad_norm": 0.1490899622440338,
      "learning_rate": 5.3735506008474e-05,
      "loss": 1.3043,
      "mean_token_accuracy": 0.7534748643636704,
      "num_tokens": 1987392447.0,
      "step": 27650
    },
    {
      "entropy": 0.8674998924136161,
      "epoch": 7.538838920686835,
      "grad_norm": 0.1464938372373581,
      "learning_rate": 5.370554518800938e-05,
      "loss": 1.2942,
      "mean_token_accuracy": 0.7565787047147751,
      "num_tokens": 1988102176.0,
      "step": 27660
    },
    {
      "entropy": 0.8758485689759254,
      "epoch": 7.541564458980648,
      "grad_norm": 0.16097642481327057,
      "learning_rate": 5.367558302962261e-05,
      "loss": 1.3015,
      "mean_token_accuracy": 0.7581014633178711,
      "num_tokens": 1988829150.0,
      "step": 27670
    },
    {
      "entropy": 0.8719006225466728,
      "epoch": 7.544289997274461,
      "grad_norm": 0.14630360901355743,
      "learning_rate": 5.364561954413179e-05,
      "loss": 1.291,
      "mean_token_accuracy": 0.758354552090168,
      "num_tokens": 1989552435.0,
      "step": 27680
    },
    {
      "entropy": 0.8725865796208382,
      "epoch": 7.547015535568275,
      "grad_norm": 0.14534975588321686,
      "learning_rate": 5.361565474235553e-05,
      "loss": 1.2947,
      "mean_token_accuracy": 0.7572773873806,
      "num_tokens": 1990277435.0,
      "step": 27690
    },
    {
      "entropy": 0.8703749105334282,
      "epoch": 7.549741073862088,
      "grad_norm": 0.15353354811668396,
      "learning_rate": 5.3585688635112905e-05,
      "loss": 1.2975,
      "mean_token_accuracy": 0.7611069425940513,
      "num_tokens": 1991010665.0,
      "step": 27700
    },
    {
      "entropy": 0.8728433132171631,
      "epoch": 7.552466612155901,
      "grad_norm": 0.14781498908996582,
      "learning_rate": 5.355572123322348e-05,
      "loss": 1.2978,
      "mean_token_accuracy": 0.7551047176122665,
      "num_tokens": 1991721417.0,
      "step": 27710
    },
    {
      "entropy": 0.8746820703148842,
      "epoch": 7.555192150449714,
      "grad_norm": 0.14716841280460358,
      "learning_rate": 5.352575254750723e-05,
      "loss": 1.3009,
      "mean_token_accuracy": 0.7551755025982857,
      "num_tokens": 1992448824.0,
      "step": 27720
    },
    {
      "entropy": 0.8664837598800659,
      "epoch": 7.557917688743527,
      "grad_norm": 0.1486971527338028,
      "learning_rate": 5.349578258878466e-05,
      "loss": 1.2968,
      "mean_token_accuracy": 0.7625233814120292,
      "num_tokens": 1993176475.0,
      "step": 27730
    },
    {
      "entropy": 0.8618240833282471,
      "epoch": 7.56064322703734,
      "grad_norm": 0.1519266813993454,
      "learning_rate": 5.34658113678767e-05,
      "loss": 1.2853,
      "mean_token_accuracy": 0.7587587416172028,
      "num_tokens": 1993886616.0,
      "step": 27740
    },
    {
      "entropy": 0.8753886893391609,
      "epoch": 7.563368765331153,
      "grad_norm": 0.14492467045783997,
      "learning_rate": 5.343583889560474e-05,
      "loss": 1.2973,
      "mean_token_accuracy": 0.7597087681293487,
      "num_tokens": 1994620813.0,
      "step": 27750
    },
    {
      "entropy": 0.8721052780747414,
      "epoch": 7.566094303624966,
      "grad_norm": 0.14899621903896332,
      "learning_rate": 5.340586518279064e-05,
      "loss": 1.3009,
      "mean_token_accuracy": 0.7571509644389153,
      "num_tokens": 1995334839.0,
      "step": 27760
    },
    {
      "entropy": 0.873625734448433,
      "epoch": 7.568819841918779,
      "grad_norm": 0.15280811488628387,
      "learning_rate": 5.337589024025668e-05,
      "loss": 1.3065,
      "mean_token_accuracy": 0.7560456842184067,
      "num_tokens": 1996042429.0,
      "step": 27770
    },
    {
      "entropy": 0.8739833995699883,
      "epoch": 7.571545380212592,
      "grad_norm": 0.15121133625507355,
      "learning_rate": 5.334591407882559e-05,
      "loss": 1.296,
      "mean_token_accuracy": 0.7578023582696914,
      "num_tokens": 1996763238.0,
      "step": 27780
    },
    {
      "entropy": 0.8746642336249352,
      "epoch": 7.574270918506405,
      "grad_norm": 0.14567768573760986,
      "learning_rate": 5.331593670932057e-05,
      "loss": 1.2883,
      "mean_token_accuracy": 0.7543858543038369,
      "num_tokens": 1997473219.0,
      "step": 27790
    },
    {
      "entropy": 0.87596046179533,
      "epoch": 7.576996456800218,
      "grad_norm": 0.15322861075401306,
      "learning_rate": 5.328595814256521e-05,
      "loss": 1.2992,
      "mean_token_accuracy": 0.7527386784553528,
      "num_tokens": 1998180186.0,
      "step": 27800
    },
    {
      "entropy": 0.8672592014074325,
      "epoch": 7.579721995094031,
      "grad_norm": 0.15870657563209534,
      "learning_rate": 5.325597838938355e-05,
      "loss": 1.2901,
      "mean_token_accuracy": 0.76070666462183,
      "num_tokens": 1998904004.0,
      "step": 27810
    },
    {
      "entropy": 0.8716690674424171,
      "epoch": 7.582447533387844,
      "grad_norm": 0.1436341106891632,
      "learning_rate": 5.3225997460600105e-05,
      "loss": 1.2896,
      "mean_token_accuracy": 0.7607816740870476,
      "num_tokens": 1999651781.0,
      "step": 27820
    },
    {
      "entropy": 0.8652073204517364,
      "epoch": 7.585173071681657,
      "grad_norm": 0.16149428486824036,
      "learning_rate": 5.319601536703972e-05,
      "loss": 1.2879,
      "mean_token_accuracy": 0.7590062454342842,
      "num_tokens": 2000364603.0,
      "step": 27830
    },
    {
      "entropy": 0.8689879328012466,
      "epoch": 7.58789860997547,
      "grad_norm": 0.148299440741539,
      "learning_rate": 5.316603211952775e-05,
      "loss": 1.2939,
      "mean_token_accuracy": 0.7555256754159927,
      "num_tokens": 2001078895.0,
      "step": 27840
    },
    {
      "entropy": 0.872168904542923,
      "epoch": 7.590624148269283,
      "grad_norm": 0.14846102893352509,
      "learning_rate": 5.3136047728889926e-05,
      "loss": 1.2955,
      "mean_token_accuracy": 0.7549246683716774,
      "num_tokens": 2001791417.0,
      "step": 27850
    },
    {
      "entropy": 0.8768918186426162,
      "epoch": 7.593349686563096,
      "grad_norm": 0.14368847012519836,
      "learning_rate": 5.310606220595238e-05,
      "loss": 1.2963,
      "mean_token_accuracy": 0.7554072514176369,
      "num_tokens": 2002515932.0,
      "step": 27860
    },
    {
      "entropy": 0.8717198982834816,
      "epoch": 7.596075224856909,
      "grad_norm": 0.14632409811019897,
      "learning_rate": 5.3076075561541675e-05,
      "loss": 1.2958,
      "mean_token_accuracy": 0.7551713183522224,
      "num_tokens": 2003222492.0,
      "step": 27870
    },
    {
      "entropy": 0.8790291532874107,
      "epoch": 7.598800763150722,
      "grad_norm": 0.1530723124742508,
      "learning_rate": 5.304608780648479e-05,
      "loss": 1.3143,
      "mean_token_accuracy": 0.7510809034109116,
      "num_tokens": 2003929619.0,
      "step": 27880
    },
    {
      "entropy": 0.8678181394934654,
      "epoch": 7.601526301444535,
      "grad_norm": 0.14335593581199646,
      "learning_rate": 5.301609895160906e-05,
      "loss": 1.2984,
      "mean_token_accuracy": 0.7595829129219055,
      "num_tokens": 2004650738.0,
      "step": 27890
    },
    {
      "entropy": 0.8779300346970558,
      "epoch": 7.604251839738348,
      "grad_norm": 0.16247470676898956,
      "learning_rate": 5.2986109007742256e-05,
      "loss": 1.2989,
      "mean_token_accuracy": 0.7545091077685356,
      "num_tokens": 2005370340.0,
      "step": 27900
    },
    {
      "entropy": 0.8720334753394127,
      "epoch": 7.606977378032162,
      "grad_norm": 0.15210209786891937,
      "learning_rate": 5.295611798571255e-05,
      "loss": 1.2955,
      "mean_token_accuracy": 0.7591853260993957,
      "num_tokens": 2006095537.0,
      "step": 27910
    },
    {
      "entropy": 0.8750460058450699,
      "epoch": 7.609702916325975,
      "grad_norm": 0.15534533560276031,
      "learning_rate": 5.2926125896348455e-05,
      "loss": 1.298,
      "mean_token_accuracy": 0.7603192970156669,
      "num_tokens": 2006829278.0,
      "step": 27920
    },
    {
      "entropy": 0.8695294588804245,
      "epoch": 7.612428454619788,
      "grad_norm": 0.15298064053058624,
      "learning_rate": 5.2896132750478914e-05,
      "loss": 1.3013,
      "mean_token_accuracy": 0.7559292703866959,
      "num_tokens": 2007537533.0,
      "step": 27930
    },
    {
      "entropy": 0.874238470196724,
      "epoch": 7.6151539929136005,
      "grad_norm": 0.150922030210495,
      "learning_rate": 5.286613855893322e-05,
      "loss": 1.296,
      "mean_token_accuracy": 0.7546189114451408,
      "num_tokens": 2008248282.0,
      "step": 27940
    },
    {
      "entropy": 0.86860870718956,
      "epoch": 7.6178795312074135,
      "grad_norm": 0.1464589536190033,
      "learning_rate": 5.28361433325411e-05,
      "loss": 1.2875,
      "mean_token_accuracy": 0.7599435091018677,
      "num_tokens": 2008974839.0,
      "step": 27950
    },
    {
      "entropy": 0.8728788539767265,
      "epoch": 7.620605069501226,
      "grad_norm": 0.1423725187778473,
      "learning_rate": 5.2806147082132575e-05,
      "loss": 1.2966,
      "mean_token_accuracy": 0.7579036712646484,
      "num_tokens": 2009694196.0,
      "step": 27960
    },
    {
      "entropy": 0.8673812478780747,
      "epoch": 7.623330607795039,
      "grad_norm": 0.1488499939441681,
      "learning_rate": 5.277614981853808e-05,
      "loss": 1.2921,
      "mean_token_accuracy": 0.7593614980578423,
      "num_tokens": 2010416044.0,
      "step": 27970
    },
    {
      "entropy": 0.8709435850381851,
      "epoch": 7.626056146088852,
      "grad_norm": 0.15130707621574402,
      "learning_rate": 5.274615155258842e-05,
      "loss": 1.293,
      "mean_token_accuracy": 0.7609807163476944,
      "num_tokens": 2011150351.0,
      "step": 27980
    },
    {
      "entropy": 0.8771456077694892,
      "epoch": 7.628781684382665,
      "grad_norm": 0.16368922591209412,
      "learning_rate": 5.271615229511475e-05,
      "loss": 1.3004,
      "mean_token_accuracy": 0.7585861459374428,
      "num_tokens": 2011875815.0,
      "step": 27990
    },
    {
      "entropy": 0.8724688425660133,
      "epoch": 7.631507222676479,
      "grad_norm": 0.15537528693675995,
      "learning_rate": 5.268615205694861e-05,
      "loss": 1.2998,
      "mean_token_accuracy": 0.7584230929613114,
      "num_tokens": 2012593558.0,
      "step": 28000
    },
    {
      "entropy": 0.8713582247495651,
      "epoch": 7.634232760970292,
      "grad_norm": 0.15075698494911194,
      "learning_rate": 5.2656150848921826e-05,
      "loss": 1.293,
      "mean_token_accuracy": 0.7569369345903396,
      "num_tokens": 2013310383.0,
      "step": 28010
    },
    {
      "entropy": 0.8728232637047768,
      "epoch": 7.636958299264105,
      "grad_norm": 0.1428200602531433,
      "learning_rate": 5.262614868186665e-05,
      "loss": 1.2967,
      "mean_token_accuracy": 0.7571837827563286,
      "num_tokens": 2014029852.0,
      "step": 28020
    },
    {
      "entropy": 0.8691148728132247,
      "epoch": 7.639683837557918,
      "grad_norm": 0.14623454213142395,
      "learning_rate": 5.259614556661564e-05,
      "loss": 1.2954,
      "mean_token_accuracy": 0.7571438029408455,
      "num_tokens": 2014749163.0,
      "step": 28030
    },
    {
      "entropy": 0.8800149992108345,
      "epoch": 7.642409375851731,
      "grad_norm": 0.1505279541015625,
      "learning_rate": 5.256614151400169e-05,
      "loss": 1.2967,
      "mean_token_accuracy": 0.7544341444969177,
      "num_tokens": 2015471406.0,
      "step": 28040
    },
    {
      "entropy": 0.8720920726656913,
      "epoch": 7.645134914145544,
      "grad_norm": 0.14527486264705658,
      "learning_rate": 5.253613653485806e-05,
      "loss": 1.3012,
      "mean_token_accuracy": 0.7559257552027703,
      "num_tokens": 2016184202.0,
      "step": 28050
    },
    {
      "entropy": 0.8746996775269509,
      "epoch": 7.647860452439357,
      "grad_norm": 0.1468832790851593,
      "learning_rate": 5.250613064001835e-05,
      "loss": 1.295,
      "mean_token_accuracy": 0.754166205227375,
      "num_tokens": 2016898886.0,
      "step": 28060
    },
    {
      "entropy": 0.8767225205898285,
      "epoch": 7.65058599073317,
      "grad_norm": 0.14846216142177582,
      "learning_rate": 5.247612384031644e-05,
      "loss": 1.3017,
      "mean_token_accuracy": 0.7529629811644554,
      "num_tokens": 2017606335.0,
      "step": 28070
    },
    {
      "entropy": 0.8717394858598709,
      "epoch": 7.653311529026983,
      "grad_norm": 0.1528032273054123,
      "learning_rate": 5.2446116146586576e-05,
      "loss": 1.2919,
      "mean_token_accuracy": 0.7587347462773323,
      "num_tokens": 2018336966.0,
      "step": 28080
    },
    {
      "entropy": 0.8684518188238144,
      "epoch": 7.656037067320796,
      "grad_norm": 0.1458507925271988,
      "learning_rate": 5.241610756966334e-05,
      "loss": 1.2909,
      "mean_token_accuracy": 0.7595139756798744,
      "num_tokens": 2019060486.0,
      "step": 28090
    },
    {
      "entropy": 0.8723010882735253,
      "epoch": 7.6587626056146085,
      "grad_norm": 0.15268127620220184,
      "learning_rate": 5.23860981203816e-05,
      "loss": 1.299,
      "mean_token_accuracy": 0.7563619449734688,
      "num_tokens": 2019775885.0,
      "step": 28100
    },
    {
      "entropy": 0.8726183950901032,
      "epoch": 7.6614881439084215,
      "grad_norm": 0.16130176186561584,
      "learning_rate": 5.2356087809576537e-05,
      "loss": 1.2984,
      "mean_token_accuracy": 0.7603875428438187,
      "num_tokens": 2020506495.0,
      "step": 28110
    },
    {
      "entropy": 0.8787395536899567,
      "epoch": 7.664213682202235,
      "grad_norm": 0.14795050024986267,
      "learning_rate": 5.2326076648083686e-05,
      "loss": 1.3029,
      "mean_token_accuracy": 0.7544093638658523,
      "num_tokens": 2021231415.0,
      "step": 28120
    },
    {
      "entropy": 0.8735887467861175,
      "epoch": 7.666939220496048,
      "grad_norm": 0.15643061697483063,
      "learning_rate": 5.229606464673884e-05,
      "loss": 1.2996,
      "mean_token_accuracy": 0.7532265886664391,
      "num_tokens": 2021946770.0,
      "step": 28130
    },
    {
      "entropy": 0.8820726230740548,
      "epoch": 7.669664758789861,
      "grad_norm": 0.14852860569953918,
      "learning_rate": 5.226605181637812e-05,
      "loss": 1.306,
      "mean_token_accuracy": 0.7542601272463798,
      "num_tokens": 2022665280.0,
      "step": 28140
    },
    {
      "entropy": 0.8707545638084412,
      "epoch": 7.672390297083674,
      "grad_norm": 0.1472589671611786,
      "learning_rate": 5.2236038167837955e-05,
      "loss": 1.2928,
      "mean_token_accuracy": 0.7565170720219612,
      "num_tokens": 2023379803.0,
      "step": 28150
    },
    {
      "entropy": 0.8759514331817627,
      "epoch": 7.675115835377487,
      "grad_norm": 0.15143978595733643,
      "learning_rate": 5.2206023711955035e-05,
      "loss": 1.2981,
      "mean_token_accuracy": 0.7590578615665435,
      "num_tokens": 2024101475.0,
      "step": 28160
    },
    {
      "entropy": 0.8768856734037399,
      "epoch": 7.6778413736713,
      "grad_norm": 0.14674776792526245,
      "learning_rate": 5.217600845956637e-05,
      "loss": 1.2945,
      "mean_token_accuracy": 0.7557491675019264,
      "num_tokens": 2024829995.0,
      "step": 28170
    },
    {
      "entropy": 0.870268528163433,
      "epoch": 7.680566911965113,
      "grad_norm": 0.14764563739299774,
      "learning_rate": 5.214599242150927e-05,
      "loss": 1.297,
      "mean_token_accuracy": 0.753543208539486,
      "num_tokens": 2025532013.0,
      "step": 28180
    },
    {
      "entropy": 0.8822046786546707,
      "epoch": 7.683292450258926,
      "grad_norm": 0.1518286168575287,
      "learning_rate": 5.2115975608621306e-05,
      "loss": 1.3045,
      "mean_token_accuracy": 0.7527895733714104,
      "num_tokens": 2026245566.0,
      "step": 28190
    },
    {
      "entropy": 0.8823674887418747,
      "epoch": 7.686017988552739,
      "grad_norm": 0.15768709778785706,
      "learning_rate": 5.20859580317403e-05,
      "loss": 1.3064,
      "mean_token_accuracy": 0.7527905866503716,
      "num_tokens": 2026958744.0,
      "step": 28200
    },
    {
      "entropy": 0.8611121058464051,
      "epoch": 7.688743526846552,
      "grad_norm": 0.15413539111614227,
      "learning_rate": 5.205593970170441e-05,
      "loss": 1.2896,
      "mean_token_accuracy": 0.7592192009091377,
      "num_tokens": 2027664655.0,
      "step": 28210
    },
    {
      "entropy": 0.8748145267367363,
      "epoch": 7.691469065140366,
      "grad_norm": 0.1530967354774475,
      "learning_rate": 5.202592062935203e-05,
      "loss": 1.3017,
      "mean_token_accuracy": 0.7531543225049973,
      "num_tokens": 2028374681.0,
      "step": 28220
    },
    {
      "entropy": 0.8767523631453514,
      "epoch": 7.694194603434179,
      "grad_norm": 0.1627414971590042,
      "learning_rate": 5.199590082552185e-05,
      "loss": 1.303,
      "mean_token_accuracy": 0.7552956163883209,
      "num_tokens": 2029088867.0,
      "step": 28230
    },
    {
      "entropy": 0.8727109178900718,
      "epoch": 7.6969201417279915,
      "grad_norm": 0.1445888876914978,
      "learning_rate": 5.1965880301052784e-05,
      "loss": 1.2971,
      "mean_token_accuracy": 0.7599951565265656,
      "num_tokens": 2029821879.0,
      "step": 28240
    },
    {
      "entropy": 0.8680927142500877,
      "epoch": 7.6996456800218045,
      "grad_norm": 0.14112494885921478,
      "learning_rate": 5.193585906678401e-05,
      "loss": 1.2839,
      "mean_token_accuracy": 0.7629461631178855,
      "num_tokens": 2030556588.0,
      "step": 28250
    },
    {
      "entropy": 0.865997676551342,
      "epoch": 7.702371218315617,
      "grad_norm": 0.16141605377197266,
      "learning_rate": 5.1905837133555013e-05,
      "loss": 1.2933,
      "mean_token_accuracy": 0.7613214045763016,
      "num_tokens": 2031286201.0,
      "step": 28260
    },
    {
      "entropy": 0.8729136109352111,
      "epoch": 7.70509675660943,
      "grad_norm": 0.14612805843353271,
      "learning_rate": 5.187581451220549e-05,
      "loss": 1.2968,
      "mean_token_accuracy": 0.7578103125095368,
      "num_tokens": 2032012448.0,
      "step": 28270
    },
    {
      "entropy": 0.867331612110138,
      "epoch": 7.707822294903243,
      "grad_norm": 0.15364515781402588,
      "learning_rate": 5.1845791213575366e-05,
      "loss": 1.2889,
      "mean_token_accuracy": 0.7628289684653282,
      "num_tokens": 2032732866.0,
      "step": 28280
    },
    {
      "entropy": 0.8800551548600197,
      "epoch": 7.710547833197056,
      "grad_norm": 0.1462222784757614,
      "learning_rate": 5.181576724850485e-05,
      "loss": 1.3009,
      "mean_token_accuracy": 0.7573458805680275,
      "num_tokens": 2033464624.0,
      "step": 28290
    },
    {
      "entropy": 0.8730786427855491,
      "epoch": 7.713273371490869,
      "grad_norm": 0.15561863780021667,
      "learning_rate": 5.178574262783441e-05,
      "loss": 1.2978,
      "mean_token_accuracy": 0.7562719061970711,
      "num_tokens": 2034181348.0,
      "step": 28300
    },
    {
      "entropy": 0.8696371629834175,
      "epoch": 7.715998909784682,
      "grad_norm": 0.13873456418514252,
      "learning_rate": 5.1755717362404676e-05,
      "loss": 1.2927,
      "mean_token_accuracy": 0.7574522286653519,
      "num_tokens": 2034894145.0,
      "step": 28310
    },
    {
      "entropy": 0.8680856406688691,
      "epoch": 7.718724448078495,
      "grad_norm": 0.15574420988559723,
      "learning_rate": 5.172569146305659e-05,
      "loss": 1.2989,
      "mean_token_accuracy": 0.7579647421836853,
      "num_tokens": 2035603297.0,
      "step": 28320
    },
    {
      "entropy": 0.8740738928318024,
      "epoch": 7.721449986372308,
      "grad_norm": 0.15564875304698944,
      "learning_rate": 5.169566494063125e-05,
      "loss": 1.296,
      "mean_token_accuracy": 0.7547544836997986,
      "num_tokens": 2036319544.0,
      "step": 28330
    },
    {
      "entropy": 0.8744974240660668,
      "epoch": 7.724175524666122,
      "grad_norm": 0.1477930098772049,
      "learning_rate": 5.1665637805970033e-05,
      "loss": 1.2962,
      "mean_token_accuracy": 0.7602969035506248,
      "num_tokens": 2037049995.0,
      "step": 28340
    },
    {
      "entropy": 0.8757786348462104,
      "epoch": 7.726901062959935,
      "grad_norm": 0.1456153392791748,
      "learning_rate": 5.1635610069914545e-05,
      "loss": 1.3061,
      "mean_token_accuracy": 0.752829161286354,
      "num_tokens": 2037765878.0,
      "step": 28350
    },
    {
      "entropy": 0.8735652223229409,
      "epoch": 7.729626601253748,
      "grad_norm": 0.14605747163295746,
      "learning_rate": 5.160558174330654e-05,
      "loss": 1.3039,
      "mean_token_accuracy": 0.7535927176475525,
      "num_tokens": 2038468813.0,
      "step": 28360
    },
    {
      "entropy": 0.8727952927350998,
      "epoch": 7.732352139547561,
      "grad_norm": 0.1495419591665268,
      "learning_rate": 5.157555283698805e-05,
      "loss": 1.2959,
      "mean_token_accuracy": 0.7583206579089165,
      "num_tokens": 2039192634.0,
      "step": 28370
    },
    {
      "entropy": 0.873136593401432,
      "epoch": 7.735077677841374,
      "grad_norm": 0.27537640929222107,
      "learning_rate": 5.154552336180132e-05,
      "loss": 1.2953,
      "mean_token_accuracy": 0.7609014913439751,
      "num_tokens": 2039915717.0,
      "step": 28380
    },
    {
      "entropy": 0.8816787675023079,
      "epoch": 7.737803216135187,
      "grad_norm": 0.1552937924861908,
      "learning_rate": 5.1515493328588725e-05,
      "loss": 1.3028,
      "mean_token_accuracy": 0.7500071495771408,
      "num_tokens": 2040616409.0,
      "step": 28390
    },
    {
      "entropy": 0.868381279706955,
      "epoch": 7.740528754429,
      "grad_norm": 0.15035855770111084,
      "learning_rate": 5.148546274819293e-05,
      "loss": 1.2919,
      "mean_token_accuracy": 0.7579583555459977,
      "num_tokens": 2041341127.0,
      "step": 28400
    },
    {
      "entropy": 0.8838419929146767,
      "epoch": 7.7432542927228125,
      "grad_norm": 0.1548195779323578,
      "learning_rate": 5.1455431631456754e-05,
      "loss": 1.3085,
      "mean_token_accuracy": 0.7517982959747315,
      "num_tokens": 2042048723.0,
      "step": 28410
    },
    {
      "entropy": 0.8783964306116104,
      "epoch": 7.7459798310166255,
      "grad_norm": 0.15469075739383698,
      "learning_rate": 5.1425399989223224e-05,
      "loss": 1.3052,
      "mean_token_accuracy": 0.7562689185142517,
      "num_tokens": 2042775977.0,
      "step": 28420
    },
    {
      "entropy": 0.8680647879838943,
      "epoch": 7.748705369310439,
      "grad_norm": 0.14739985764026642,
      "learning_rate": 5.139536783233552e-05,
      "loss": 1.3002,
      "mean_token_accuracy": 0.7525867044925689,
      "num_tokens": 2043468264.0,
      "step": 28430
    },
    {
      "entropy": 0.8728204682469368,
      "epoch": 7.751430907604252,
      "grad_norm": 0.14950290322303772,
      "learning_rate": 5.1365335171637054e-05,
      "loss": 1.2964,
      "mean_token_accuracy": 0.7546725392341613,
      "num_tokens": 2044191301.0,
      "step": 28440
    },
    {
      "entropy": 0.8709087863564491,
      "epoch": 7.754156445898065,
      "grad_norm": 0.14465369284152985,
      "learning_rate": 5.13353020179714e-05,
      "loss": 1.2959,
      "mean_token_accuracy": 0.7577947944402694,
      "num_tokens": 2044922129.0,
      "step": 28450
    },
    {
      "entropy": 0.8761092722415924,
      "epoch": 7.756881984191878,
      "grad_norm": 0.14949002861976624,
      "learning_rate": 5.130526838218233e-05,
      "loss": 1.297,
      "mean_token_accuracy": 0.7524799570441246,
      "num_tokens": 2045627395.0,
      "step": 28460
    },
    {
      "entropy": 0.8751310005784034,
      "epoch": 7.759607522485691,
      "grad_norm": 0.14488691091537476,
      "learning_rate": 5.1275234275113746e-05,
      "loss": 1.2875,
      "mean_token_accuracy": 0.7553243070840836,
      "num_tokens": 2046348000.0,
      "step": 28470
    },
    {
      "entropy": 0.8674445793032646,
      "epoch": 7.762333060779504,
      "grad_norm": 0.1486712098121643,
      "learning_rate": 5.124519970760976e-05,
      "loss": 1.2806,
      "mean_token_accuracy": 0.7652259364724159,
      "num_tokens": 2047087265.0,
      "step": 28480
    },
    {
      "entropy": 0.8799532428383827,
      "epoch": 7.765058599073317,
      "grad_norm": 0.15229366719722748,
      "learning_rate": 5.121516469051463e-05,
      "loss": 1.298,
      "mean_token_accuracy": 0.7570998996496201,
      "num_tokens": 2047812014.0,
      "step": 28490
    },
    {
      "entropy": 0.8761812463402748,
      "epoch": 7.76778413736713,
      "grad_norm": 0.15411978960037231,
      "learning_rate": 5.11851292346728e-05,
      "loss": 1.2991,
      "mean_token_accuracy": 0.7579398721456527,
      "num_tokens": 2048543242.0,
      "step": 28500
    },
    {
      "entropy": 0.8693125024437904,
      "epoch": 7.770509675660943,
      "grad_norm": 0.14534714818000793,
      "learning_rate": 5.115509335092884e-05,
      "loss": 1.2988,
      "mean_token_accuracy": 0.7580270186066628,
      "num_tokens": 2049259454.0,
      "step": 28510
    },
    {
      "entropy": 0.8759208023548126,
      "epoch": 7.773235213954756,
      "grad_norm": 0.15540142357349396,
      "learning_rate": 5.112505705012749e-05,
      "loss": 1.2991,
      "mean_token_accuracy": 0.7556175932288169,
      "num_tokens": 2049983695.0,
      "step": 28520
    },
    {
      "entropy": 0.8707679629325866,
      "epoch": 7.775960752248569,
      "grad_norm": 0.1550641804933548,
      "learning_rate": 5.109502034311365e-05,
      "loss": 1.2924,
      "mean_token_accuracy": 0.7534662991762161,
      "num_tokens": 2050682950.0,
      "step": 28530
    },
    {
      "entropy": 0.8711427360773086,
      "epoch": 7.778686290542382,
      "grad_norm": 0.1522579938173294,
      "learning_rate": 5.106498324073235e-05,
      "loss": 1.2936,
      "mean_token_accuracy": 0.7548329576849937,
      "num_tokens": 2051406021.0,
      "step": 28540
    },
    {
      "entropy": 0.8687419131398201,
      "epoch": 7.7814118288361955,
      "grad_norm": 0.1569044142961502,
      "learning_rate": 5.103494575382875e-05,
      "loss": 1.2891,
      "mean_token_accuracy": 0.7605527967214585,
      "num_tokens": 2052138302.0,
      "step": 28550
    },
    {
      "entropy": 0.8768089443445206,
      "epoch": 7.7841373671300085,
      "grad_norm": 0.14389604330062866,
      "learning_rate": 5.100490789324822e-05,
      "loss": 1.2981,
      "mean_token_accuracy": 0.7566765040159226,
      "num_tokens": 2052856325.0,
      "step": 28560
    },
    {
      "entropy": 0.8717021763324737,
      "epoch": 7.786862905423821,
      "grad_norm": 0.1500464230775833,
      "learning_rate": 5.0974869669836176e-05,
      "loss": 1.3067,
      "mean_token_accuracy": 0.7527828842401505,
      "num_tokens": 2053547861.0,
      "step": 28570
    },
    {
      "entropy": 0.880061000585556,
      "epoch": 7.789588443717634,
      "grad_norm": 0.16395455598831177,
      "learning_rate": 5.09448310944382e-05,
      "loss": 1.3128,
      "mean_token_accuracy": 0.7522459611296654,
      "num_tokens": 2054247397.0,
      "step": 28580
    },
    {
      "entropy": 0.8706887930631637,
      "epoch": 7.792313982011447,
      "grad_norm": 0.152640700340271,
      "learning_rate": 5.091479217790004e-05,
      "loss": 1.3015,
      "mean_token_accuracy": 0.7574447631835938,
      "num_tokens": 2054956031.0,
      "step": 28590
    },
    {
      "entropy": 0.8678671464323997,
      "epoch": 7.79503952030526,
      "grad_norm": 0.1524311751127243,
      "learning_rate": 5.088475293106748e-05,
      "loss": 1.2858,
      "mean_token_accuracy": 0.7643247053027153,
      "num_tokens": 2055696865.0,
      "step": 28600
    },
    {
      "entropy": 0.8778502061963082,
      "epoch": 7.797765058599073,
      "grad_norm": 0.16135413944721222,
      "learning_rate": 5.085471336478651e-05,
      "loss": 1.2953,
      "mean_token_accuracy": 0.7575853750109672,
      "num_tokens": 2056422012.0,
      "step": 28610
    },
    {
      "entropy": 0.8740970060229302,
      "epoch": 7.800490596892886,
      "grad_norm": 0.1481577605009079,
      "learning_rate": 5.08246734899032e-05,
      "loss": 1.3054,
      "mean_token_accuracy": 0.7566355332732201,
      "num_tokens": 2057143210.0,
      "step": 28620
    },
    {
      "entropy": 0.8765804812312126,
      "epoch": 7.803216135186699,
      "grad_norm": 0.14563089609146118,
      "learning_rate": 5.0794633317263705e-05,
      "loss": 1.2969,
      "mean_token_accuracy": 0.7538089320063591,
      "num_tokens": 2057848448.0,
      "step": 28630
    },
    {
      "entropy": 0.8724101155996322,
      "epoch": 7.805941673480512,
      "grad_norm": 0.14638759195804596,
      "learning_rate": 5.076459285771434e-05,
      "loss": 1.2959,
      "mean_token_accuracy": 0.7598433643579483,
      "num_tokens": 2058577055.0,
      "step": 28640
    },
    {
      "entropy": 0.871101850271225,
      "epoch": 7.808667211774326,
      "grad_norm": 0.14928314089775085,
      "learning_rate": 5.073455212210149e-05,
      "loss": 1.2905,
      "mean_token_accuracy": 0.7581345871090889,
      "num_tokens": 2059297531.0,
      "step": 28650
    },
    {
      "entropy": 0.8746340870857239,
      "epoch": 7.811392750068139,
      "grad_norm": 0.15390898287296295,
      "learning_rate": 5.070451112127164e-05,
      "loss": 1.3031,
      "mean_token_accuracy": 0.7531252816319466,
      "num_tokens": 2059994462.0,
      "step": 28660
    },
    {
      "entropy": 0.8774883389472962,
      "epoch": 7.814118288361952,
      "grad_norm": 0.1507336050271988,
      "learning_rate": 5.067446986607138e-05,
      "loss": 1.3043,
      "mean_token_accuracy": 0.7546640068292618,
      "num_tokens": 2060717996.0,
      "step": 28670
    },
    {
      "entropy": 0.8740352123975754,
      "epoch": 7.816843826655765,
      "grad_norm": 0.1557072103023529,
      "learning_rate": 5.0644428367347384e-05,
      "loss": 1.3012,
      "mean_token_accuracy": 0.7583262786269188,
      "num_tokens": 2061442048.0,
      "step": 28680
    },
    {
      "entropy": 0.8714375674724579,
      "epoch": 7.819569364949578,
      "grad_norm": 0.14534524083137512,
      "learning_rate": 5.0614386635946444e-05,
      "loss": 1.2973,
      "mean_token_accuracy": 0.7554130285978318,
      "num_tokens": 2062152326.0,
      "step": 28690
    },
    {
      "entropy": 0.8720107138156891,
      "epoch": 7.822294903243391,
      "grad_norm": 0.15054894983768463,
      "learning_rate": 5.0584344682715376e-05,
      "loss": 1.2911,
      "mean_token_accuracy": 0.7564399510622024,
      "num_tokens": 2062863326.0,
      "step": 28700
    },
    {
      "entropy": 0.8656201481819152,
      "epoch": 7.8250204415372036,
      "grad_norm": 0.15560075640678406,
      "learning_rate": 5.055430251850115e-05,
      "loss": 1.2894,
      "mean_token_accuracy": 0.7588586747646332,
      "num_tokens": 2063574513.0,
      "step": 28710
    },
    {
      "entropy": 0.8750212281942368,
      "epoch": 7.8277459798310165,
      "grad_norm": 0.14586029946804047,
      "learning_rate": 5.052426015415074e-05,
      "loss": 1.2966,
      "mean_token_accuracy": 0.7538041993975639,
      "num_tokens": 2064290500.0,
      "step": 28720
    },
    {
      "entropy": 0.8781249269843101,
      "epoch": 7.8304715181248294,
      "grad_norm": 0.15518252551555634,
      "learning_rate": 5.049421760051124e-05,
      "loss": 1.3055,
      "mean_token_accuracy": 0.7511297702789307,
      "num_tokens": 2064992594.0,
      "step": 28730
    },
    {
      "entropy": 0.8679031729698181,
      "epoch": 7.833197056418642,
      "grad_norm": 0.1644304394721985,
      "learning_rate": 5.046417486842981e-05,
      "loss": 1.2928,
      "mean_token_accuracy": 0.7591112449765205,
      "num_tokens": 2065704189.0,
      "step": 28740
    },
    {
      "entropy": 0.8719989255070686,
      "epoch": 7.835922594712455,
      "grad_norm": 0.1510053128004074,
      "learning_rate": 5.043413196875364e-05,
      "loss": 1.3027,
      "mean_token_accuracy": 0.7548772737383842,
      "num_tokens": 2066429087.0,
      "step": 28750
    },
    {
      "entropy": 0.8725510746240616,
      "epoch": 7.838648133006268,
      "grad_norm": 0.16147853434085846,
      "learning_rate": 5.0404088912330014e-05,
      "loss": 1.2959,
      "mean_token_accuracy": 0.7553753957152367,
      "num_tokens": 2067146968.0,
      "step": 28760
    },
    {
      "entropy": 0.869031710922718,
      "epoch": 7.841373671300082,
      "grad_norm": 0.15349683165550232,
      "learning_rate": 5.0374045710006265e-05,
      "loss": 1.2946,
      "mean_token_accuracy": 0.7622701898217201,
      "num_tokens": 2067874528.0,
      "step": 28770
    },
    {
      "entropy": 0.8651412725448608,
      "epoch": 7.844099209593895,
      "grad_norm": 0.15462660789489746,
      "learning_rate": 5.034400237262977e-05,
      "loss": 1.2952,
      "mean_token_accuracy": 0.7564776301383972,
      "num_tokens": 2068591259.0,
      "step": 28780
    },
    {
      "entropy": 0.879278439283371,
      "epoch": 7.846824747887708,
      "grad_norm": 0.15315796434879303,
      "learning_rate": 5.031395891104796e-05,
      "loss": 1.2998,
      "mean_token_accuracy": 0.7557676270604133,
      "num_tokens": 2069315225.0,
      "step": 28790
    },
    {
      "entropy": 0.8675184726715088,
      "epoch": 7.849550286181521,
      "grad_norm": 0.16138049960136414,
      "learning_rate": 5.0283915336108334e-05,
      "loss": 1.2907,
      "mean_token_accuracy": 0.7601470157504082,
      "num_tokens": 2070031797.0,
      "step": 28800
    },
    {
      "entropy": 0.8679433673620224,
      "epoch": 7.852275824475334,
      "grad_norm": 0.15830840170383453,
      "learning_rate": 5.025387165865837e-05,
      "loss": 1.2881,
      "mean_token_accuracy": 0.7589368507266044,
      "num_tokens": 2070744676.0,
      "step": 28810
    },
    {
      "entropy": 0.8728272631764412,
      "epoch": 7.855001362769147,
      "grad_norm": 0.17641285061836243,
      "learning_rate": 5.022382788954564e-05,
      "loss": 1.2967,
      "mean_token_accuracy": 0.7578941285610199,
      "num_tokens": 2071463729.0,
      "step": 28820
    },
    {
      "entropy": 0.8773080214858056,
      "epoch": 7.85772690106296,
      "grad_norm": 0.14152438938617706,
      "learning_rate": 5.019378403961774e-05,
      "loss": 1.3003,
      "mean_token_accuracy": 0.7510676220059395,
      "num_tokens": 2072164198.0,
      "step": 28830
    },
    {
      "entropy": 0.8713888987898827,
      "epoch": 7.860452439356773,
      "grad_norm": 0.14663468301296234,
      "learning_rate": 5.016374011972227e-05,
      "loss": 1.2947,
      "mean_token_accuracy": 0.7552718207240104,
      "num_tokens": 2072874975.0,
      "step": 28840
    },
    {
      "entropy": 0.8713116094470024,
      "epoch": 7.863177977650586,
      "grad_norm": 0.1441587209701538,
      "learning_rate": 5.013369614070689e-05,
      "loss": 1.2916,
      "mean_token_accuracy": 0.7578077897429466,
      "num_tokens": 2073587049.0,
      "step": 28850
    },
    {
      "entropy": 0.8819049805402756,
      "epoch": 7.8659035159443995,
      "grad_norm": 0.14788268506526947,
      "learning_rate": 5.010365211341925e-05,
      "loss": 1.3095,
      "mean_token_accuracy": 0.7508459910750389,
      "num_tokens": 2074295657.0,
      "step": 28860
    },
    {
      "entropy": 0.8664075076580048,
      "epoch": 7.8686290542382125,
      "grad_norm": 0.1527957022190094,
      "learning_rate": 5.007360804870702e-05,
      "loss": 1.2945,
      "mean_token_accuracy": 0.7563967630267143,
      "num_tokens": 2075002325.0,
      "step": 28870
    },
    {
      "entropy": 0.8703516647219658,
      "epoch": 7.871354592532025,
      "grad_norm": 0.14753733575344086,
      "learning_rate": 5.004356395741792e-05,
      "loss": 1.2959,
      "mean_token_accuracy": 0.7564651280641556,
      "num_tokens": 2075714080.0,
      "step": 28880
    },
    {
      "entropy": 0.8724898785352707,
      "epoch": 7.874080130825838,
      "grad_norm": 0.15951234102249146,
      "learning_rate": 5.0013519850399636e-05,
      "loss": 1.2997,
      "mean_token_accuracy": 0.7589197918772698,
      "num_tokens": 2076433153.0,
      "step": 28890
    },
    {
      "entropy": 0.8677097305655479,
      "epoch": 7.876805669119651,
      "grad_norm": 0.15661266446113586,
      "learning_rate": 4.99834757384999e-05,
      "loss": 1.2907,
      "mean_token_accuracy": 0.7581684812903404,
      "num_tokens": 2077155366.0,
      "step": 28900
    },
    {
      "entropy": 0.8718841284513473,
      "epoch": 7.879531207413464,
      "grad_norm": 0.15242601931095123,
      "learning_rate": 4.995343163256638e-05,
      "loss": 1.299,
      "mean_token_accuracy": 0.7597081452608109,
      "num_tokens": 2077877506.0,
      "step": 28910
    },
    {
      "entropy": 0.8715924382209778,
      "epoch": 7.882256745707277,
      "grad_norm": 0.14559628069400787,
      "learning_rate": 4.992338754344681e-05,
      "loss": 1.2964,
      "mean_token_accuracy": 0.7535214453935624,
      "num_tokens": 2078576137.0,
      "step": 28920
    },
    {
      "entropy": 0.8692012846469879,
      "epoch": 7.88498228400109,
      "grad_norm": 0.1578289419412613,
      "learning_rate": 4.9893343481988916e-05,
      "loss": 1.2982,
      "mean_token_accuracy": 0.7584845826029778,
      "num_tokens": 2079284511.0,
      "step": 28930
    },
    {
      "entropy": 0.865885728597641,
      "epoch": 7.887707822294903,
      "grad_norm": 0.15666401386260986,
      "learning_rate": 4.986329945904034e-05,
      "loss": 1.2919,
      "mean_token_accuracy": 0.7594642907381057,
      "num_tokens": 2080011636.0,
      "step": 28940
    },
    {
      "entropy": 0.8837577745318412,
      "epoch": 7.890433360588716,
      "grad_norm": 0.15608501434326172,
      "learning_rate": 4.983325548544881e-05,
      "loss": 1.3067,
      "mean_token_accuracy": 0.748767901957035,
      "num_tokens": 2080709720.0,
      "step": 28950
    },
    {
      "entropy": 0.8710357174277306,
      "epoch": 7.893158898882529,
      "grad_norm": 0.15831336379051208,
      "learning_rate": 4.9803211572061964e-05,
      "loss": 1.2969,
      "mean_token_accuracy": 0.7558223530650139,
      "num_tokens": 2081422223.0,
      "step": 28960
    },
    {
      "entropy": 0.8780325442552567,
      "epoch": 7.895884437176342,
      "grad_norm": 0.14892467856407166,
      "learning_rate": 4.977316772972742e-05,
      "loss": 1.3054,
      "mean_token_accuracy": 0.7538846135139465,
      "num_tokens": 2082140287.0,
      "step": 28970
    },
    {
      "entropy": 0.8729042440652848,
      "epoch": 7.898609975470155,
      "grad_norm": 0.15484024584293365,
      "learning_rate": 4.974312396929284e-05,
      "loss": 1.2995,
      "mean_token_accuracy": 0.7566973924636841,
      "num_tokens": 2082852070.0,
      "step": 28980
    },
    {
      "entropy": 0.8725142627954483,
      "epoch": 7.901335513763969,
      "grad_norm": 0.15064504742622375,
      "learning_rate": 4.971308030160578e-05,
      "loss": 1.2988,
      "mean_token_accuracy": 0.756444227695465,
      "num_tokens": 2083569938.0,
      "step": 28990
    },
    {
      "entropy": 0.8781291499733925,
      "epoch": 7.904061052057782,
      "grad_norm": 0.14941169321537018,
      "learning_rate": 4.9683036737513766e-05,
      "loss": 1.3013,
      "mean_token_accuracy": 0.7528211146593093,
      "num_tokens": 2084280826.0,
      "step": 29000
    },
    {
      "entropy": 0.8703803390264511,
      "epoch": 7.906786590351595,
      "grad_norm": 0.15744654834270477,
      "learning_rate": 4.9652993287864365e-05,
      "loss": 1.307,
      "mean_token_accuracy": 0.7544958233833313,
      "num_tokens": 2084986611.0,
      "step": 29010
    },
    {
      "entropy": 0.874736875295639,
      "epoch": 7.9095121286454075,
      "grad_norm": 0.15729069709777832,
      "learning_rate": 4.962294996350502e-05,
      "loss": 1.3026,
      "mean_token_accuracy": 0.7613536983728408,
      "num_tokens": 2085720092.0,
      "step": 29020
    },
    {
      "entropy": 0.8789043530821801,
      "epoch": 7.9122376669392205,
      "grad_norm": 0.15260839462280273,
      "learning_rate": 4.959290677528314e-05,
      "loss": 1.3005,
      "mean_token_accuracy": 0.7594448283314705,
      "num_tokens": 2086458948.0,
      "step": 29030
    },
    {
      "entropy": 0.8734215199947357,
      "epoch": 7.914963205233033,
      "grad_norm": 0.1524142473936081,
      "learning_rate": 4.9562863734046136e-05,
      "loss": 1.2991,
      "mean_token_accuracy": 0.7579223662614822,
      "num_tokens": 2087182747.0,
      "step": 29040
    },
    {
      "entropy": 0.8781965866684913,
      "epoch": 7.917688743526846,
      "grad_norm": 0.15536931157112122,
      "learning_rate": 4.9532820850641324e-05,
      "loss": 1.3014,
      "mean_token_accuracy": 0.7552846103906632,
      "num_tokens": 2087901479.0,
      "step": 29050
    },
    {
      "entropy": 0.8621720016002655,
      "epoch": 7.920414281820659,
      "grad_norm": 0.15621359646320343,
      "learning_rate": 4.9502778135915946e-05,
      "loss": 1.2844,
      "mean_token_accuracy": 0.7571163326501846,
      "num_tokens": 2088614224.0,
      "step": 29060
    },
    {
      "entropy": 0.8729408875107765,
      "epoch": 7.923139820114472,
      "grad_norm": 0.1571531891822815,
      "learning_rate": 4.9472735600717244e-05,
      "loss": 1.3032,
      "mean_token_accuracy": 0.7551058560609818,
      "num_tokens": 2089319357.0,
      "step": 29070
    },
    {
      "entropy": 0.8744441866874695,
      "epoch": 7.925865358408286,
      "grad_norm": 0.15197701752185822,
      "learning_rate": 4.944269325589235e-05,
      "loss": 1.2983,
      "mean_token_accuracy": 0.7555972620844841,
      "num_tokens": 2090024117.0,
      "step": 29080
    },
    {
      "entropy": 0.8694946125149727,
      "epoch": 7.928590896702099,
      "grad_norm": 0.14163386821746826,
      "learning_rate": 4.94126511122883e-05,
      "loss": 1.2849,
      "mean_token_accuracy": 0.7586724862456322,
      "num_tokens": 2090754500.0,
      "step": 29090
    },
    {
      "entropy": 0.8758626118302345,
      "epoch": 7.931316434995912,
      "grad_norm": 0.15822726488113403,
      "learning_rate": 4.938260918075215e-05,
      "loss": 1.2957,
      "mean_token_accuracy": 0.7548936754465103,
      "num_tokens": 2091479909.0,
      "step": 29100
    },
    {
      "entropy": 0.8677125558257103,
      "epoch": 7.934041973289725,
      "grad_norm": 0.15288494527339935,
      "learning_rate": 4.935256747213079e-05,
      "loss": 1.2901,
      "mean_token_accuracy": 0.7586245208978653,
      "num_tokens": 2092201401.0,
      "step": 29110
    },
    {
      "entropy": 0.8700149923563003,
      "epoch": 7.936767511583538,
      "grad_norm": 0.16054970026016235,
      "learning_rate": 4.932252599727107e-05,
      "loss": 1.295,
      "mean_token_accuracy": 0.7596103355288506,
      "num_tokens": 2092922481.0,
      "step": 29120
    },
    {
      "entropy": 0.8743564888834954,
      "epoch": 7.939493049877351,
      "grad_norm": 0.15474098920822144,
      "learning_rate": 4.929248476701975e-05,
      "loss": 1.2959,
      "mean_token_accuracy": 0.7529025077819824,
      "num_tokens": 2093630352.0,
      "step": 29130
    },
    {
      "entropy": 0.8784813895821572,
      "epoch": 7.942218588171164,
      "grad_norm": 0.14834600687026978,
      "learning_rate": 4.9262443792223514e-05,
      "loss": 1.3037,
      "mean_token_accuracy": 0.7554646402597427,
      "num_tokens": 2094361932.0,
      "step": 29140
    },
    {
      "entropy": 0.8759296759963036,
      "epoch": 7.944944126464977,
      "grad_norm": 0.1595781147480011,
      "learning_rate": 4.923240308372891e-05,
      "loss": 1.2998,
      "mean_token_accuracy": 0.7544375360012054,
      "num_tokens": 2095070081.0,
      "step": 29150
    },
    {
      "entropy": 0.8672859281301498,
      "epoch": 7.94766966475879,
      "grad_norm": 0.15980692207813263,
      "learning_rate": 4.920236265238244e-05,
      "loss": 1.2935,
      "mean_token_accuracy": 0.7607377246022224,
      "num_tokens": 2095793054.0,
      "step": 29160
    },
    {
      "entropy": 0.877199088037014,
      "epoch": 7.950395203052603,
      "grad_norm": 0.1558620184659958,
      "learning_rate": 4.91723225090305e-05,
      "loss": 1.3007,
      "mean_token_accuracy": 0.7527604401111603,
      "num_tokens": 2096502020.0,
      "step": 29170
    },
    {
      "entropy": 0.8671931162476539,
      "epoch": 7.953120741346416,
      "grad_norm": 0.14660750329494476,
      "learning_rate": 4.9142282664519325e-05,
      "loss": 1.2905,
      "mean_token_accuracy": 0.7587891101837159,
      "num_tokens": 2097227198.0,
      "step": 29180
    },
    {
      "entropy": 0.8784502819180489,
      "epoch": 7.9558462796402285,
      "grad_norm": 0.15164071321487427,
      "learning_rate": 4.9112243129695135e-05,
      "loss": 1.2952,
      "mean_token_accuracy": 0.7544587805867196,
      "num_tokens": 2097955629.0,
      "step": 29190
    },
    {
      "entropy": 0.8754410952329635,
      "epoch": 7.958571817934042,
      "grad_norm": 0.1389971375465393,
      "learning_rate": 4.908220391540396e-05,
      "loss": 1.2894,
      "mean_token_accuracy": 0.7591995671391487,
      "num_tokens": 2098695455.0,
      "step": 29200
    },
    {
      "entropy": 0.8730250328779221,
      "epoch": 7.961297356227855,
      "grad_norm": 0.1592153012752533,
      "learning_rate": 4.905216503249173e-05,
      "loss": 1.3056,
      "mean_token_accuracy": 0.7555223658680916,
      "num_tokens": 2099404174.0,
      "step": 29210
    },
    {
      "entropy": 0.8666514158248901,
      "epoch": 7.964022894521668,
      "grad_norm": 0.15044082701206207,
      "learning_rate": 4.902212649180431e-05,
      "loss": 1.2927,
      "mean_token_accuracy": 0.7587355688214302,
      "num_tokens": 2100128215.0,
      "step": 29220
    },
    {
      "entropy": 0.8808118417859078,
      "epoch": 7.966748432815481,
      "grad_norm": 0.14613182842731476,
      "learning_rate": 4.8992088304187375e-05,
      "loss": 1.3007,
      "mean_token_accuracy": 0.7540887221693993,
      "num_tokens": 2100831599.0,
      "step": 29230
    },
    {
      "entropy": 0.869738195836544,
      "epoch": 7.969473971109294,
      "grad_norm": 0.15028458833694458,
      "learning_rate": 4.896205048048647e-05,
      "loss": 1.2982,
      "mean_token_accuracy": 0.7567011937499046,
      "num_tokens": 2101555506.0,
      "step": 29240
    },
    {
      "entropy": 0.8647333770990372,
      "epoch": 7.972199509403107,
      "grad_norm": 0.15632669627666473,
      "learning_rate": 4.893201303154709e-05,
      "loss": 1.2968,
      "mean_token_accuracy": 0.7580927804112434,
      "num_tokens": 2102264508.0,
      "step": 29250
    },
    {
      "entropy": 0.8750887006521225,
      "epoch": 7.97492504769692,
      "grad_norm": 0.14812743663787842,
      "learning_rate": 4.890197596821451e-05,
      "loss": 1.3058,
      "mean_token_accuracy": 0.7583898589015007,
      "num_tokens": 2102994858.0,
      "step": 29260
    },
    {
      "entropy": 0.8756055474281311,
      "epoch": 7.977650585990733,
      "grad_norm": 0.14465123414993286,
      "learning_rate": 4.887193930133388e-05,
      "loss": 1.2929,
      "mean_token_accuracy": 0.7579392105340957,
      "num_tokens": 2103725694.0,
      "step": 29270
    },
    {
      "entropy": 0.879104483127594,
      "epoch": 7.980376124284546,
      "grad_norm": 0.15725469589233398,
      "learning_rate": 4.884190304175025e-05,
      "loss": 1.3039,
      "mean_token_accuracy": 0.7550291955471039,
      "num_tokens": 2104446339.0,
      "step": 29280
    },
    {
      "entropy": 0.8693584635853767,
      "epoch": 7.983101662578359,
      "grad_norm": 0.16086307168006897,
      "learning_rate": 4.88118672003085e-05,
      "loss": 1.2962,
      "mean_token_accuracy": 0.7578431040048599,
      "num_tokens": 2105158797.0,
      "step": 29290
    },
    {
      "entropy": 0.8693640872836113,
      "epoch": 7.985827200872173,
      "grad_norm": 0.14272333681583405,
      "learning_rate": 4.878183178785332e-05,
      "loss": 1.2951,
      "mean_token_accuracy": 0.7567166343331337,
      "num_tokens": 2105866549.0,
      "step": 29300
    },
    {
      "entropy": 0.874030065536499,
      "epoch": 7.988552739165986,
      "grad_norm": 0.1451587677001953,
      "learning_rate": 4.8751796815229325e-05,
      "loss": 1.2994,
      "mean_token_accuracy": 0.756797869503498,
      "num_tokens": 2106602225.0,
      "step": 29310
    },
    {
      "entropy": 0.8713746085762978,
      "epoch": 7.991278277459799,
      "grad_norm": 0.1409689337015152,
      "learning_rate": 4.872176229328089e-05,
      "loss": 1.2997,
      "mean_token_accuracy": 0.7568163946270943,
      "num_tokens": 2107314621.0,
      "step": 29320
    },
    {
      "entropy": 0.8717377915978431,
      "epoch": 7.9940038157536115,
      "grad_norm": 0.14872410893440247,
      "learning_rate": 4.869172823285227e-05,
      "loss": 1.3008,
      "mean_token_accuracy": 0.7556920111179352,
      "num_tokens": 2108022731.0,
      "step": 29330
    },
    {
      "entropy": 0.8727119475603103,
      "epoch": 7.9967293540474245,
      "grad_norm": 0.1454230099916458,
      "learning_rate": 4.866169464478757e-05,
      "loss": 1.297,
      "mean_token_accuracy": 0.7566224813461304,
      "num_tokens": 2108738992.0,
      "step": 29340
    },
    {
      "entropy": 0.8854518592357635,
      "epoch": 7.999454892341237,
      "grad_norm": 0.14412358403205872,
      "learning_rate": 4.863166153993067e-05,
      "loss": 1.3024,
      "mean_token_accuracy": 0.753192912042141,
      "num_tokens": 2109461047.0,
      "step": 29350
    }
  ],
  "logging_steps": 10,
  "max_steps": 55035,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 15,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 2.050092967582478e+20,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}