view article Article Tokenmaxxing BERT: 50x faster tokenization huggingface • about 21 hours ago • 40
UltraBERT Collection UltraBERT: Beyond the Limit of BERT Pre-Training. Pre-trained 150M/380M encoders and a text embedding model, English, 32K context. • 4 items • Updated 2 days ago • 27
mrm8488/embeddinggemma-2-ft-typed-decision Text Classification • 0.3B • Updated 2 days ago • 925 • 60
Running Agents 39 The Arabic Information Retrieval (AIR) Leaderboard 📊 39 Evaluating No, Full & Late Interaction Models 🤠
JEV-as-a-Judge: Accept When Confident, Escalate When Unsure Paper • 2609.26550 • Published 24 days ago • 42
view article Article mLateOn: A New SoTA for Multilingual ColBERT-Style Retrieval, Evaluated on HAKARI-Bench hotchpotch • Aug 13 • 10
view article Article Introducing jev-reranker: Reranking and Relevance Filtering for RAG hotchpotch • 27 days ago • 10