paper-with-me

Papers

Revisiting Multi-Agent Debate as Test-Time Scaling: A Systematic Study of Conditional Effectiveness

2025-05-29 · Yongjin Yang, Euiin Yi, Jongwoo Ko, Kimin Lee, Zhijing Jin, Se-Young Yun

The remarkable growth in large language model (LLM) capabilities has spurred exploration into multi-agent systems, with debate frameworks emerging as a promising avenue for enhanced problem-solving. These multi-agent debate (MAD) approaches, where agents collaboratively present, critique, and refine arguments, potentially offer improved reasoning, robustness, and diverse perspectives over monolithic models. Despite prior studies leveraging MAD, a systematic understanding of its effectiveness compared to self-agent methods, particularly under varying conditions, remains elusive. This paper seeks to fill this gap by conceptualizing MAD as a test-time computational scaling technique, distinguished by collaborative refinement and diverse exploration capabilities. We conduct a comprehensive empirical investigation comparing MAD with strong self-agent test-time scaling baselines on mathematical reasoning and safety-related tasks. Our study systematically examines the influence of task difficulty, model scale, and agent diversity on MAD's performance. Key findings reveal that, for mathematical reasoning, MAD offers limited advantages over self-agent scaling but becomes more effective with increased problem difficulty and decreased model capability, while agent diversity shows little benefit. Conversely, for safety tasks, MAD's collaborative refinement can increase vulnerability, but incorporating diverse agent configurations facilitates a gradual reduction in attack success through the collaborative refinement process. We believe our findings provide critical guidance for the future development of more effective and strategically deployed MAD systems.

📄 PDF Abstract BibTeX arXiv:2505.22960

Code (0)

등록된 구현이 없습니다.

Tasks

DiversityLarge Language ModelMathematical Reasoning

Similar Papers 제목 키워드 기반

Learning from Self-Debate: Preparing Reasoning Models for Multi-Agent Debate

2026-01-29 · Chenxi Liu, Yanshuo Chen, Ruibo Chen, Tianyi Xiong 외 arxiv

The reasoning abilities of large language models (LLMs) have been substantially improved by reinforcement learning with verifiable rewards (RLVR). At test time, collaborative reasoning through Multi-Agent Debate (MAD) ha…

Reinforcement Learning

Stay Focused: Problem Drift in Multi-Agent Debate

2025-02-26 · Jonas Becker, Lars Benedikt Kaesberg, Andreas Stephan, Jan Philip Wahle 외

Multi-agent debate - multiple instances of large language models discussing problems in turn-based interaction - has shown promise for solving knowledge and reasoning tasks. However, these methods show limitations when s…

Instruction Following

An LLM-Driven Multi-Agent Debate System for Mendelian Diseases

2025-04-10 · Xinyang Zhou, Yongyong Ren, Qianqian Zhao, Daoyi Huang 외

Accurate diagnosis of Mendelian diseases is crucial for precision therapy and assistance in preimplantation genetic diagnosis. However, existing methods often fall short of clinical standards or depend on extensive datas…

DiagnosticLanguage ModelingLanguage Modelling

MALLM: Multi-Agent Large Language Models Framework

2025-09-15 · Jonas Becker, Lars Benedikt Kaesberg, Niklas Bauer, Jan Philip Wahle 외 arxiv

Multi-agent debate (MAD) has demonstrated the ability to augment collective intelligence by scaling test-time compute and leveraging expertise. Current frameworks for multi-agent debate are often designed towards tool us…

Multi-Agent Debate for LLM Judges with Adaptive Stability Detection

2025-10-14 · Tianyu Hu, Zhen Tan, Song Wang, Huaizhi Qu 외 arxiv

With advancements in reasoning capabilities, Large Language Models (LLMs) are increasingly employed for automated judgment tasks. While LLMs-as-Judges offer promise in automating evaluations, current approaches often rel…

Computational Efficiency