Improving LLM Reasoning with Multi-Agent Tree-of-Thought Validator Agent

TL;DR

This paper introduces a multi-agent Tree-of-Thought approach with a Thought Validator to improve reasoning accuracy and trustworthiness in large language models, outperforming existing methods on the GSM8K dataset.

Contribution

It presents a novel combination of ToT reasoning with a validator agent to filter out flawed reasoning paths, enhancing robustness and accuracy.

Findings

Outperforms standard ToT by 5.6% on GSM8K

Enables more trustworthy reasoning through validation

Demonstrates effectiveness across four different LLMs

Abstract

Multi-agent strategies have emerged as a promising approach to enhance the reasoning abilities of Large Language Models (LLMs) by assigning specialized roles in the problem-solving process. Concurrently, Tree of Thoughts (ToT) methods have shown potential in improving reasoning for complex question-answering tasks by exploring diverse reasoning paths. A critical limitation in multi-agent reasoning is the 'Reasoner' agent's shallow exploration of reasoning paths. While ToT strategies could help mitigate this problem, they may generate flawed reasoning branches, which could harm the trustworthiness of the final answer. To leverage the strengths of both multi-agent reasoning and ToT strategies, we introduce a novel approach combining ToT-based Reasoner agents with a Thought Validator agent. Multiple Reasoner agents operate in parallel, employing ToT to explore diverse reasoning paths. The Thought Validator then scrutinizes these paths, considering a Reasoner's conclusion only if its reasoning is valid. This method enables a more robust voting strategy by discarding faulty reasoning paths, enhancing the system's ability to tackle tasks requiring systematic and trustworthy reasoning. Our method demonstrates superior performance compared to existing techniques when evaluated on the GSM8K dataset, outperforming the standard ToT strategy by an average 5.6% across four LLMs. The code and related content can be found in: https://github.com/SecureAIAutonomyLab/MA-ToT