view article Article Improving Hugging Face Training Efficiency Through Packing with Flash Attention 17 days ago • 16
Understanding Reference Policies in Direct Preference Optimization Paper • 2407.13709 • Published Jul 18 • 15
view article Article RegMix: Data Mixture as Regression for Language Model Pre-training By SivilTaram • Jul 11 • 8
Paloma Collection Dataset and baseline models for Paloma, a benchmark of language model fit to 585 textual domains • 8 items • Updated 9 days ago • 13