幻影共识：测量多智能体LLM决策中的社会推理
原标题：Phantom Consensus: Measuring Social Reasoning in Multi-Agent LLM Decision-Making
AI 摘要
该论文提出「幻影共识」（phantom consensus）概念，指多智能体LLM系统在观察到模糊社会信号后，公开决策趋同于一种群体私下并不支持的共识，并错误估计他人的私下判断。作者构建PHANTOM-Bench基准，分离私下判断、公开决策、群体私下信念估计、透明恢复及对照条件，在九个闭源与开源模型上测试。结果显示模糊公开信号可使公开准确率最多下降47.8个百分点，最易受影响场景下集体理性失败率达47.3%，而公开记录的真实私下决策可在不泄露金标准标签的情况下修复多数失败。
正文节选
SocialAgent: Second Workshop on Large Language Models for Social Reasoning and Simulation Phantom Consensus: Measuring Social Reasoning in Multi-Agent LLM Decision-Making Abstract Multi-agent LLM systems increasingly aggregate judgments across model instances, but reasoning about other agents can itself introduce failure. We study a setting in which LLMs shift their public decisions after observing ambiguous social signals while reporting inaccurate estimates of what other agents privately judge