PolyPythias
- Preview • Updated • 52
EleutherAI/pile-preshuffled-seeds
Updated • 329 • 1Note Training data information for each seed.
-
EleutherAI/pythia-14m-deduped
Text Generation • 39.2M • Updated • 65.4k • 28 -
EleutherAI/pythia-14m-seed1
Text Generation • Updated • 176 -
EleutherAI/pythia-14m-seed2
Text Generation • Updated • 233 -
EleutherAI/pythia-14m-seed3
Text Generation • Updated • 93 -
EleutherAI/pythia-14m-seed4
Text Generation • Updated • 93 -
EleutherAI/pythia-14m-seed5
Text Generation • Updated • 116 -
EleutherAI/pythia-14m-seed6
Text Generation • Updated • 100 -
EleutherAI/pythia-14m-seed7
Text Generation • Updated • 105 -
EleutherAI/pythia-14m-seed8
Text Generation • Updated • 100 -
EleutherAI/pythia-14m-seed9
Text Generation • Updated • 100 -
EleutherAI/pythia-31m-deduped
Text Generation • 55.7M • Updated • 3.48k • 5 -
EleutherAI/pythia-31m-seed1
Text Generation • Updated • 196 -
EleutherAI/pythia-31m-seed2
Text Generation • Updated • 196 -
EleutherAI/pythia-31m-seed3
Text Generation • Updated • 83 -
EleutherAI/pythia-31m-seed4
Text Generation • Updated • 78 -
EleutherAI/pythia-31m-seed5
Text Generation • Updated • 66 -
EleutherAI/pythia-31m-seed6
Text Generation • Updated • 66 -
EleutherAI/pythia-31m-seed7
Text Generation • Updated • 71 -
EleutherAI/pythia-31m-seed8
Text Generation • Updated • 69 -
EleutherAI/pythia-31m-seed9
Text Generation • Updated • 68 -
EleutherAI/pythia-70m
95.6M • Updated • 174k • 79 -
EleutherAI/pythia-70m-seed1
Text Generation • Updated • 999 -
EleutherAI/pythia-70m-seed2
Text Generation • Updated • 675 -
EleutherAI/pythia-70m-seed3
Text Generation • Updated • 597 -
EleutherAI/pythia-70m-seed4
Text Generation • Updated • 567 -
EleutherAI/pythia-70m-seed5
Text Generation • Updated • 555 -
EleutherAI/pythia-70m-seed6
Text Generation • Updated • 528 -
EleutherAI/pythia-70m-seed7
Text Generation • Updated • 559 -
EleutherAI/pythia-70m-seed8
Text Generation • Updated • 537 -
EleutherAI/pythia-70m-seed9
Text Generation • Updated • 522 -
EleutherAI/pythia-160m
Text Generation • Updated • 2.52M • 38 -
EleutherAI/pythia-160m-seed1
Text Generation • 0.2B • Updated • 1.5k -
EleutherAI/pythia-160m-seed2
Text Generation • 0.2B • Updated • 1.4k -
EleutherAI/pythia-160m-seed3
Text Generation • 0.2B • Updated • 1.29k -
EleutherAI/pythia-160m-seed4
Text Generation • Updated • 910 • 1 -
EleutherAI/pythia-160m-seed5
Text Generation • Updated • 638 -
EleutherAI/pythia-160m-seed6
Text Generation • Updated • 617 -
EleutherAI/pythia-160m-seed7
Text Generation • Updated • 623 -
EleutherAI/pythia-160m-seed8
Text Generation • Updated • 611 -
EleutherAI/pythia-160m-seed9
Text Generation • Updated • 613 -
EleutherAI/pythia-410m
Text Generation • 0.5B • Updated • 90.5k • 36 -
EleutherAI/pythia-410m-seed1
Text Generation • Updated • 678 -
EleutherAI/pythia-410m-seed2
Text Generation • Updated • 1.33k -
EleutherAI/pythia-410m-seed3
Text Generation • Updated • 624 -
EleutherAI/pythia-410m-seed4
Text Generation • Updated • 569 -
EleutherAI/pythia-410m-seed5
Text Generation • Updated • 548 -
EleutherAI/pythia-410m-seed6
Text Generation • Updated • 890 • 1 -
EleutherAI/pythia-410m-seed7
Text Generation • Updated • 590 -
EleutherAI/pythia-410m-seed8
Text Generation • Updated • 555 -
EleutherAI/pythia-410m-seed9
Text Generation • Updated • 746
EleutherAI/pythia-160m-data-seed1
Text Generation • Updated • 124Note Version where the data order and weight initialization seeds are decoupled. Here, only the data seed is different from pythia-160m ("seed 0").
EleutherAI/pythia-160m-data-seed2
Text Generation • Updated • 106Note Version where the data order and weight initialization seeds are decoupled. Here, only the data seed is different from pythia-160m ("seed 0").
EleutherAI/pythia-160m-data-seed3
Text Generation • Updated • 105Note Version where the data order and weight initialization seeds are decoupled. Here, only the data seed is different from pythia-160m ("seed 0").
EleutherAI/pythia-160m-weight-seed1
Text Generation • Updated • 264Note Version where the data order and weight initialization seeds are decoupled. Here, only the weight initialization seed is different from pythia-160m ("seed 0").
EleutherAI/pythia-160m-weight-seed2
Text Generation • Updated • 211Note Version where the data order and weight initialization seeds are decoupled. Here, only the weight initialization seed is different from pythia-160m ("seed 0").
EleutherAI/pythia-160m-weight-seed3
Text Generation • Updated • 201Note Version where the data order and weight initialization seeds are decoupled. Here, only the weight initialization seed is different from pythia-160m ("seed 0").
-
PolyPythias: Stability and Outliers across Fifty Language Model Pre-Training Runs
Paper • 2503.09543 • Published