用推理树审计多智能体逻辑，比投票和AI判官更准
Auditing Multi-Agent LLM Reasoning Trees Outperforms Majority Vote and LLM-as-Judge
- 构建推理树，分析智能体间逻辑分歧点
- 在5个场景中比投票高5%准确率，比AI判官高3%
- 适合追求高可靠推理的AI系统开发者
多智能体系统可显著提升大语言模型的推理能力，但现有框架多采用多数投票聚合结果。该方法忽略推理过程中的证据结构，在共谋共识（即智能体共享偏差并趋向相同错误推论）时表现脆弱。本文提出AgentAuditor，通过显式表示智能体推理路径间的共识与分歧，以路径搜索替代投票；在关键分歧点对比推理分支，将全局裁决转化为高效局部验证。进一步提出抗共识偏好优化（ACPO），让裁判模型在多数失败案例中学习，奖励基于证据的少数正确选择而非流行错误。AgentAuditor不依赖具体多智能体设置，在5种主流设置下均实现：相比多数投票绝对提升最高5%，相比LLM-as-Judge最高提升3%。
原文摘要 · Abstract (English)
Multi-agent systems (MAS) can substantially extend the reasoning capacity of large language models (LLMs), yet most frameworks still aggregate agent outputs with majority voting. This heuristic discards the evidential structure of reasoning traces and is brittle under the confabulation consensus, where agents share correlated biases and converge on the same incorrect rationale. We introduce AgentAuditor, which replaces voting with a path search over a Reasoning Tree that explicitly represents agreements and divergences among agent traces. AgentAuditor resolves conflicts by comparing reasoning branches at critical divergence points, turning global adjudication into efficient, localized verification. We further propose Anti-Consensus Preference Optimization (ACPO), which trains the adjudicator on majority-failure cases and rewards evidence-based minority selections over popular errors. AgentAuditor is agnostic to MAS setting, and we find across 5 popular settings that it yields up to 5% absolute accuracy improvement over a majority vote, and up to 3% over using LLM-as-Judge.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。