Diversity of Thought Elicits Stronger Reasoning Capabilities in Multi-Agent Debate Frameworks
For researchers and practitioners using LLMs for reasoning tasks, this work demonstrates that diversity among debating agents yields stronger performance than homogeneous groups, even surpassing larger models.
The paper shows that using diverse models in multi-agent debate improves reasoning accuracy, with a diverse set of medium-capacity models (Gemini-Pro, Mixtral 7BX8, PaLM 2-M) achieving 91% on GSM-8K and 94% on ASDiv, outperforming GPT-4.
Large language models (LLMs) excel in natural language generation but often confidently produce incorrect responses, especially in tasks like mathematical reasoning. Chain-of-thought prompting, self-verification, and multi-agent debate are among the strategies proposed to improve the reasoning and factual accuracy of LLMs. Building on Du et al.'s multi-agent debate framework, we find that multi-agent debate helps at any model scale, and that diversity of thought elicits stronger reasoning in debating LLMs. Across various model sizes, performance on mathematical reasoning tasks benefits most when diverse trained models are used. Remarkably, after 4 rounds of debate, a diverse set of medium-capacity models (Gemini-Pro, Mixtral 7BX8, and PaLM 2-M) outperforms GPT-4 on the GSM-8K benchmark, scoring 91% accuracy. By comparison, when 3 instances of Gemini-Pro are used, performance only reaches 82%. Finally, this diverse set of medium-capacity models sets a new state-of-the-art performance on the ASDiv benchmark (94%). These results underscore the idea that the future of AI is agentic, with diverse cooperating agents yielding emergent capabilities beyond even the most powerful individual models.