English(EN)Emergence of Biased Consensus in Multi-Agent LLM Debates
新研究应对LLM辩论挑战，引入基准和本地化协议 · 跟踪6个来源
作者PulseAugur 编辑部·[8 个来源]·
研究人员正在通过多智能体辩论（MAD）框架探索提高大型语言模型（LLM）推理能力的方法。两篇论文解决了这些辩论中LLM的“盲目从众”问题，提出了诸如规范辩论关系（DEAR）或分析由噪声驱动的偏见共识的出现等解决方案。另一项研究引入了M3MAD-Bench，这是一个用于评估跨不同领域和模态的MAD的综合基准，并强调协作方法更鲁棒但效率较低。最后，提出了一种本地化多智能体辩论（LMAD）协议，该协议将辩论限制在冲突片段内，以提高问答任务的效率和准确性。
AI
arXiv:2608.01463v2 Announce Type: replace Abstract: Multi-agent debate commonly exchanges complete rationales even when disagreements concern only a few intermediate claims. We introduce Localized Multi-Agent Debate (LMAD), an inference-time protocol that represents agent rationa…
Multi-Agent Debate (MAD) improves the reasoning performance of Large Language Models (LLMs) through multi-round interaction. However, LLMs in MAD are highly susceptible to blind conformity. Existing individual evaluation methods, typically based on confidence or perplexity, fail …
Multi-agent LLM debates achieve strong performance on decision-making tasks as well as problem-solving benchmarks, yet their safety and fairness risks remain poorly understood. Notably, interaction can amplify the biases of single LLMs, raising concerns for real-world deployment.…
arXiv cs.AI
TIER_1English(EN)·Ao Li, Jinghui Zhang, Luyu Li, Yuxiang Duan, Lang Gao, Mingcai Chen, Weijun Qin, Shaopeng Li, Fengxian Ji, Ning Liu, Lizhen Cui, Xiuying Chen, Yuntao Du·
arXiv:2601.02854v2 Announce Type: replace Abstract: As an agent-level reasoning and coordination paradigm, Multi-Agent Debate (MAD) orchestrates multiple agents through structured debate to improve answer quality and support complex reasoning. However, existing research on MAD su…
Multi-agent debate commonly exchanges complete reasoning traces even when disagreements concern only a few intermediate claims. We introduce Localized Multi-Agent Debate (LMAD), an inference-time protocol that represents agent traces as typed nodes, locates their earliest conflic…
Multi-agent debate commonly exchanges complete rationales even when disagreements concern only a few intermediate claims. We introduce Localized Multi-Agent Debate (LMAD), an inference-time protocol that represents agent rationales as nodes, locates their earliest conflict, and r…
arXiv cs.AI
TIER_1English(EN)·Quim Motger, Marc Oriol, Jordi Marco, Xavier Franch·
arXiv:2607.26212v1 Announce Type: cross Abstract: Multi-Agent Debate (MAD) is a promising paradigm for improving the accuracy and robustness of Large Language Model (LLM)-based agentic systems. It enables multiple agents to exchange arguments, critique each other's outputs, and i…
dev.to — LLM tag
TIER_1English(EN)·Devanshu Biswas·
<p>Ask one model the same question three times and you get three paraphrases and a confident tone — whether or not the answer is actually contested. A multi-agent debate does the opposite. Three personas answer independently, a critic names each one's flaw, every agent revises or…
