Recent progress in LLMs discussion suggests that multi-agent discussion improves the reasoning abilities of LLMs. In this work, we reevaluate this claim through systematic experiments, where we propose a novel group discussion framework to enrich the set of discussion mechanisms. Interestingly, our results show that a single-agent LLM with strong prompts can achieve almost the same performance as the best existing discussion approach on a wide range of reasoning tasks and backbone LLMs. We observe that the multi-agent discussion performs better than a single agent only when there is no demonstration in the prompt. Further study reveals the common interaction mechanisms of LLMs during the discussion.
@article{arxiv.2402.18272,
title = {Rethinking the Bounds of LLM Reasoning: Are Multi-Agent Discussions the Key?},
author = {Qineng Wang and Zihao Wang and Ying Su and Hanghang Tong and Yangqiu Song},
journal= {arXiv preprint arXiv:2402.18272},
year = {2024}
}