MAR:Multi-Agent Reflexion Improves Reasoning Abilities in LLMs

TL;DR

This paper introduces a multi-agent debating approach with diverse personas to enhance reasoning in LLMs, overcoming degeneration from repeated reflections and achieving superior accuracy on reasoning benchmarks.

Contribution

The novel multi-agent debating framework with multiple personas improves reflection diversity and reasoning accuracy in LLMs compared to single-agent methods.

Findings

Achieved 47% EM HotPot QA accuracy

Reached 82.7% on HumanEval programming tasks

Outperformed single-LLM reflection methods

Abstract

LLMs have shown the capacity to improve their performance on reasoning tasks through reflecting on their mistakes, and acting with these reflections in mind. However, continual reflections of the same LLM onto itself exhibit degeneration of thought, where the LLM continues to repeat the same errors again and again even with the knowledge that its wrong. To address this problem, we instead introduce multi-agent with multi-persona debators as the method to generate reflections. Through out extensive experimentation, we've found that the leads to better diversity of in the reflections generated by the llm agent. We demonstrate an accuracy of 47% EM HotPot QA (question answering) and 82.7% on HumanEval (programming), both performances surpassing reflection with a single llm.