Embedding Domain Knowledge for Large Language Models via Reinforcement Learning from Augmented Generation

Chaojun Nie, Jun Zhou, Guanxiang Wang +2

Large language models (LLMs) often exhibit limited performance on domain-specific tasks due to the natural disproportionate representation of specialized information in their train…