Improving Factuality and Reasoning in Language Models through Multiagent Debate
Yilun Du MIT CSAIL、Shuang Li MIT CSAIL、Antonio Torralba MIT CSAIL、Joshua B. Tenenbaum MIT CSAIL, BCS, CBMM、Igor Mordatch Google Brain
Abstract
Large language models (LLMs) have demonstrated remarkable capabilities in language generation, understanding, and few-shot learning in recent years. An extensive body of work has explored how their performance may be further improved through the tools of prompting, ranging from verification, self-consistency, or intermediate scratchpads. In this paper, we present a complementary approach to improve language responses where multiple language model instances propose and debate their individual responses and reasoning processes over multiple rounds to arrive at a common final answer. Our findings indicate that this approach significantly enhances mathematical and strategic reasoning across a number of tasks. We also demonstrate that our approach improves the factual validity of generated content, reducing fallacious answers and hallucinations that contemporary models are prone to. Our approach may be directly applied to existing black-box models and uses identical procedure and prompts for all tasks we investigate. Overall, our findings suggest that such "society of minds" approach has the potential to significantly advance the capabilities of LLMs and pave the way for further breakth
中文速览
让多个大语言模型（LLM）互相"辩论"来提升答题准确率——这正是本文要解决的核心问题：单个模型容易自信地"幻想"错误事实、或在推理链中出现跳跃，而现有的提示工程手段都只在单个模型上做文章。研究者提出"多智能体辩论"（Multiagent Debate）框架：让多个同类或不同类的LLM实例各自先独立作答，再把彼此的回答拼接成上下文，互相批判并更新自己的答案，如此迭代若干轮后汇聚出最终共识答案。在算术计算、小学数学（GSM8K）、国际象棋走法预测、人物传记事实核查、MMLU知识问答等六项任务上，该方法显著优于单模型基线、自我反思和多数投票等方案，尤其值得注意的是，即便所有模型一开始都答错，辩论过程本身也能引导它们最终找到正确答案。这项工作的意义在于：它无需访问模型内部权重或梯度，可直接套用在任何黑盒API上，并与链式思维等现有提示技巧正交叠加，为通过"群体智慧"大幅提升LLM可靠性提供了一条开箱即用的新路径。
原文 arXiv:2305.14325;中英对照 + 大白话阅读 https://aha.fim.ai/paper/2305.14325v1