Direct Nash Optimization aligns language models with general preferences
DNO-Prct iterates on-policy sampling, rating, and contrastive updates to raise length-controlled win rates on AlpacaEval 2.0 for Llama 3.1 8B and Mistral 7B Instruct models.
Stories per day · 30 days
Relatedalpacaeval1dataset1direct-nash-optimization1huggingface1msafetemp1multilingual-safety1preference-alignment1template-augmentation1
3 stories, newest first · one card per story
Stories tagged #llm-alignment
New MSafeTemp dataset advances multilingual LLM safety alignment
The template-based pipeline produces a 60-language corpus for safety training, exposing gaps in English-only evaluation.
LexReward: A Taxonomy-Driven Reward Framework for Legal Language Models
LexReward defines taxonomy-driven rewards over Style, Element, and Chain dimensions for legal LLMs, feeding DPO and per-dimension reward models for RL.