arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2605.02307cs.MA
收藏

SOTOPIA-TOM:基于心智理论的多智能体交互中隐私与信息管理评估

SOTOPIA-TOM: Evaluating Privacy and Information Management in Multi-Agent Interaction with Theory of Mind

Language Technologies Institute, Carnegie Mellon University(卡内基梅隆大学语言技术研究所);Fujitsu Research of America(富士通美国研究院)

Yashwanth YS, Ruichen Wang, Shihua Zeng, Xuhui Zhou, Koichi Onoue, Vasudha Varadarajan, Maarten Sap

AI总结:

本研究推出SOTOPIA-TOM基准框架,评估LLM智能体在多智能体交互中的隐私与信息管理能力,发现现有模型存在缺陷,基于ToM的干预可改善协调-隐私平衡。

摘要:

随着基于大语言模型(LLM)的智能体在多方场景中交互日益频繁,它们需要妥善处理信息不对称问题,即知晓何时以及向谁披露信息是合适的。然而,现有基准无法在现实多方场景中衡量这一能力。因此,我们推出SOTOPIA-TOM,这是一个多维基准框架,用于评估LLM智能体成功应对信息不对称且隐私敏感的多方交互的能力。我们构建了一个交互环境,支持公开(广播)和私有(直接消息)通信,并精心设计了覆盖8个行业领域的160个人工审核场景,每个场景涉及3至5个智能体,拥有分区的私有知识和依赖渠道的共享策略。为衡量交互能力,我们创建了一个多维评估框架,评估智能体共享有用信息、寻求缺失细节、高效协调以及保护隐私的能力,还将这些能力整合为综合INFOMGMT指标。结果显示,在6种LLM主干和提示策略(vanilla、CoT-privacy以及基于ToM的干预措施)中,即使是最大的高推理模型GPT-5,INFOMGMT得分也仅为62%,这表明在信息寻求和隐私感知决策方面仍存在持续的缺陷。此外,基于ToM的干预措施能更稳定地提升整体协调-隐私平衡,例如,相较于vanilla基线,ToM-Coach将GPT-4o的关键隐私违规率从9.9%降至2.2%,同时将综合InfoMgmt得分从15%提高了2.5倍以上至40%。总体而言,SOTOPIA-TOM揭示了当前LLM智能体在复杂、信息不对称协调方面的持续局限性,并为开发更具隐私感知、具备心智理论能力的多智能体系统提供了可扩展的测试平台。

英文摘要:

As LLM-based agents are increasingly interacting in multi-party settings, they need to properly handle information asymmetry, i.e., knowing when and to whom to disclose information is appropriate. Yet, existing benchmarks fail to measure this ability in realistic multi-party settings. Thus, we introduce SOTOPIA-TOM, a multi-dimensional benchmarking framework to evaluate LLM agents' ability to successfully navigate information asymmetric and privacy sensitive multi-party interactions. We create an interaction environment which enables both public (broadcast) and private (direct message) communication, and craft 160 human-reviewed scenarios across eight industry sectors, each involving 3 to 5 agents with partitioned private knowledge and channel-dependent sharing policies. To measure interaction abilities, we create a multi-dimensional evaluation framework to assess how well agents share useful information, seek missing details, coordinate efficiently, and protect privacy, which we also combine into a composite INFOMGMT metric. Results show that, across 6 LLM backbones and prompting strategies (vanilla, CoT-privacy, and ToM-based interventions), even the largest high-reasoning model (GPT-5) reaches only a 62% INFOMGMT score, which indicates persistent deficiencies in information seeking and privacy-aware decision-making. Additionally, ToM-based interventions more consistently improve the overall coordination-privacy balance (for example, relative to the vanilla baseline, ToM-Coach reduces critical privacy violations on GPT-4o from 9.9% to 2.2% while increasing the composite InfoMgmt score more than 2.5x from 15% to 40%). Overall, SOTOPIA-TOM exposes persistent limitations of current LLM agents in complex, information-asymmetric coordination and provides an extensible testbed for developing more privacy-aware, theory-of-mind capable multi-agent systems.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑