ECPEC

LLM conversation analysis: Pair judgment outperforms generation

A new paper from arXiv explores the effectiveness of different approaches for emotion-cause pair extraction in conversations using large language models (LLMs). The research found that formulating the task as a pair-lev…