Stop Overvaluing Multi-Agent Debate -- We Must Rethink Evaluation and Embrace Model Heterogeneity

• URL: http://arxiv.org/abs/2502.08788v3
• Date: Sat, 21 Jun 2025 09:22:22 GMT
• Title: Stop Overvaluing Multi-Agent Debate -- We Must Rethink Evaluation and Embrace Model Heterogeneity
• Authors: Hangfan Zhang, Zhiyao Cui, Jianhao Chen, Xinrun Wang, Qiaosheng Zhang, Zhen Wang, Dinghao Wu, Shuyue Hu,
• Abstract summary: Multi-agent debate (MAD) has gained significant attention as a promising line of research to improve the factual accuracy and reasoning capabilities of large language models (LLMs)<n>Despite its conceptual appeal, current MAD research suffers from critical limitations in evaluation practices.<n>This paper presents a systematic evaluation of 5 representative MAD methods across 9 benchmarks using 4 foundational models.
• Score: 20.408720462383158
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Multi-agent debate (MAD) has gained significant attention as a promising line of research to improve the factual accuracy and reasoning capabilities of large language models (LLMs). Despite its conceptual appeal, current MAD research suffers from critical limitations in evaluation practices, including limited benchmark coverage, weak baseline comparisons, and inconsistent setups. This paper presents a systematic evaluation of 5 representative MAD methods across 9 benchmarks using 4 foundational models. Surprisingly, our findings reveal that MAD often fail to outperform simple single-agent baselines such as Chain-of-Thought and Self-Consistency, even when consuming significantly more inference-time computation. To advance MAD research, we further explore the role of model heterogeneity and find it as a universal antidote to consistently improve current MAD frameworks. Based on our findings, we argue that the field must stop overvaluing MAD in its current form; for true advancement, we must critically rethink evaluation paradigms and actively embrace model heterogeneity as a core design principle.

Related papers

• Multi-Agent Debate Strategies to Enhance Requirements Engineering with Large Language Models [3.4829662575293585]
  Large Language Model (LLM) agents are becoming widely used for various Requirements Engineering (RE) tasks.<n>Research on improving their accuracy mainly focuses on prompt engineering, model fine-tuning, and retrieval augmented generation.<n>We propose that just as human debates enhance accuracy and reduce bias in RE tasks by incorporating diverse perspectives, different LLM agents debating and collaborating may achieve similar improvements.
  arXiv  Detail & Related papers  (2025-07-08T13:37:59Z)
• Reasoning Models Are More Easily Gaslighted Than You Think [85.84943447589511]
  We evaluate three state-of-the-art reasoning models, including OpenAI's o4-mini, Claude-3.7-Sonnet and Gemini-2.5-Flash.<n>Our evaluation reveals significant accuracy drops following gaslighting negation prompts.<n>We introduce GaslightingBench-R, a new diagnostic benchmark designed to evaluate reasoning models' susceptibility to defend their belief.
  arXiv  Detail & Related papers  (2025-06-11T12:52:25Z)
• Revisiting Multi-Agent Debate as Test-Time Scaling: A Systematic Study of Conditional Effectiveness [50.29739337771454]
  Multi-agent debate (MAD) approaches offer improved reasoning, robustness, and diverse perspectives over monolithic models.<n>This paper conceptualizes MAD as a test-time computational scaling technique, distinguished by collaborative refinement and diverse exploration capabilities.<n>We conduct a comprehensive empirical investigation comparing MAD with strong self-agent test-time scaling baselines on mathematical reasoning and safety-related tasks.
  arXiv  Detail & Related papers  (2025-05-29T01:02:55Z)
• Model Utility Law: Evaluating LLMs beyond Performance through Mechanism Interpretable Metric [99.56567010306807]
  Large Language Models (LLMs) have become indispensable across academia, industry, and daily applications.<n>One core challenge of evaluation in the large language model (LLM) era is the generalization issue.<n>We propose Model Utilization Index (MUI), a mechanism interpretability enhanced metric that complements traditional performance scores.
  arXiv  Detail & Related papers  (2025-04-10T04:09:47Z)
• MAMM-Refine: A Recipe for Improving Faithfulness in Generation with Multi-Agent Collaboration [63.31211701741323]
  We extend multi-agent multi-model reasoning to generation, specifically to improving faithfulness through refinement. We design intrinsic evaluations for each subtask, with our findings indicating that both multi-agent (multiple instances) and multi-model (diverse LLM types) approaches benefit error detection and critiquing. We consolidate these insights into a final "recipe" called Multi-Agent Multi-Model Refinement (MAMM-Refine), where multi-agent and multi-model collaboration significantly boosts performance.
  arXiv  Detail & Related papers  (2025-03-19T14:46:53Z)
• Is Multi-Agent Debate (MAD) the Silver Bullet? An Empirical Analysis of MAD in Code Summarization and Translation [10.038721196640864]
  Multi-Agent Debate (MAD) systems enable structured debates among Large Language Models (LLMs) MAD promotes divergent thinking through role-specific agents, dynamic interactions, and structured decision-making. This study investigates MAD's effectiveness on two Software Engineering (SE) tasks.
  arXiv  Detail & Related papers  (2025-03-15T07:30:37Z)
• Adversarial Alignment for LLMs Requires Simpler, Reproducible, and More Measurable Objectives [52.863024096759816]
  Misaligned research objectives have hindered progress in adversarial robustness research over the past decade.<n>We argue that realigned objectives are necessary for meaningful progress in adversarial alignment.
  arXiv  Detail & Related papers  (2025-02-17T15:28:40Z)
• LlaMADRS: Prompting Large Language Models for Interview-Based Depression Assessment [75.44934940580112]
  This study introduces LlaMADRS, a novel framework leveraging open-source Large Language Models (LLMs) to automate depression severity assessment.<n>We employ a zero-shot prompting strategy with carefully designed cues to guide the model in interpreting and scoring transcribed clinical interviews.<n>Our approach, tested on 236 real-world interviews, demonstrates strong correlations with clinician assessments.
  arXiv  Detail & Related papers  (2025-01-07T08:49:04Z)
• M-MAD: Multidimensional Multi-Agent Debate Framework for Fine-grained Machine Translation Evaluation [12.042804590050089]
  Multidimensional Multi-Agent Debate (M-MAD) is a systematic LLM-based multi-agent framework for advanced machine translation evaluation.<n>Our findings demonstrate that M-MAD achieves significant advancements by (1) decoupling MQM criteria into distinct evaluation dimensions for fine-grained assessments.<n> Comprehensive experiments show that M-MAD not only outperforms all existing LLM-as-a-judge methods but also competes with state-of-the-art reference-based automatic metrics.
  arXiv  Detail & Related papers  (2024-12-28T12:11:28Z)
• A NotSo Simple Way to Beat Simple Bench [0.0]
  This paper presents a novel framework for enhancing reasoning capabilities in large language models (LLMs)<n>We propose a multi-step prompting strategy coupled with global consistency checks to improve model accuracy and robustness.<n>Our results reveal model-specific strengths: Claude excels in maintaining logical consistency, while GPT-4o exhibits exploratory creativity but struggles with ambiguous prompts.
  arXiv  Detail & Related papers  (2024-12-12T16:04:31Z)
• MR-Ben: A Meta-Reasoning Benchmark for Evaluating System-2 Thinking in LLMs [55.20845457594977]
  Large language models (LLMs) have shown increasing capability in problem-solving and decision-making. We present a process-based benchmark MR-Ben that demands a meta-reasoning skill. Our meta-reasoning paradigm is especially suited for system-2 slow thinking.
  arXiv  Detail & Related papers  (2024-06-20T03:50:23Z)
• Look Before You Decide: Prompting Active Deduction of MLLMs for Assumptive Reasoning [77.72128397088409]
  We show that most prevalent MLLMs can be easily fooled by the introduction of a presupposition into the question. We also propose a novel reinforcement learning paradigm to encourage the model to actively perform composite deduction.
  arXiv  Detail & Related papers  (2024-04-19T15:53:27Z)
• Large Multimodal Agents: A Survey [78.81459893884737]
  Large language models (LLMs) have achieved superior performance in powering text-based AI agents. There is an emerging research trend focused on extending these LLM-powered AI agents into the multimodal domain. This review aims to provide valuable insights and guidelines for future research in this rapidly evolving field.
  arXiv  Detail & Related papers  (2024-02-23T06:04:23Z)
• Should we be going MAD? A Look at Multi-Agent Debate Strategies for LLMs [7.7433783185451075]
  We benchmark a range of debating and prompting strategies to explore the trade-offs between cost, time, and accuracy. We find that multi-agent debating systems, in their current form, do not reliably outperform other proposed prompting strategies. We build on these results to offer insights into improving debating strategies, such as adjusting agent agreement levels.
  arXiv  Detail & Related papers  (2023-11-29T05:54:41Z)
• Understanding, Predicting and Better Resolving Q-Value Divergence in Offline-RL [86.0987896274354]
  We first identify a fundamental pattern, self-excitation, as the primary cause of Q-value estimation divergence in offline RL. We then propose a novel Self-Excite Eigenvalue Measure (SEEM) metric to measure the evolving property of Q-network at training. For the first time, our theory can reliably decide whether the training will diverge at an early stage.
  arXiv  Detail & Related papers  (2023-10-06T17:57:44Z)
• Towards Characterizing Domain Counterfactuals For Invertible Latent Causal Models [15.817239008727789]
  In this work, we analyze a specific type of causal query called domain counterfactuals, which hypothesizes what a sample would have looked like if it had been generated in a different domain. We show that recovering the latent Structural Causal Model (SCM) is unnecessary for estimating domain counterfactuals. We also develop a theoretically grounded practical algorithm that simplifies the modeling process to generative model estimation.
  arXiv  Detail & Related papers  (2023-06-20T04:19:06Z)
• Encouraging Divergent Thinking in Large Language Models through Multi-Agent Debate [85.3444184685235]
  We propose a Multi-Agent Debate (MAD) framework, in which multiple agents express their arguments in the state of "tit for tat" and a judge manages the debate process to obtain a final solution. Our framework encourages divergent thinking in LLMs which would be helpful for tasks that require deep levels of contemplation.
  arXiv  Detail & Related papers  (2023-05-30T15:25:45Z)
• VERITE: A Robust Benchmark for Multimodal Misinformation Detection Accounting for Unimodal Bias [17.107961913114778]
  multimodal misinformation is a growing problem on social media platforms. In this study, we investigate and identify the presence of unimodal bias in widely-used MMD benchmarks. We introduce a new method -- termed Crossmodal HArd Synthetic MisAlignment (CHASMA) -- for generating realistic synthetic training data.
  arXiv  Detail & Related papers  (2023-04-27T12:28:29Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.