用多智能体分解隐私任务，减少信息泄露
1-2-3 Check: Enhancing Contextual Privacy in LLM via Multi-Agent Reasoning
- 拆分隐私识别为提取、分类等子任务，降低单个智能体负担
- 在ConfAIde和PrivacyLens上，私密信息泄露减少18%~19%
- 适合关注对话系统隐私保护的研究者与开发者
在交互式场景中，大型语言模型（LLMs）需处理来自多个来源的信息（如包含私密与公开内容的会议总结），这给上下文隐私保护带来挑战。本文提出一种多智能体框架，将隐私推理分解为专业化子任务（提取、分类），减轻单个智能体的信息负荷，同时支持迭代验证，提升对上下文隐私规范的遵守可靠性。通过系统性消融实验分析信息流拓扑，揭示上游检测错误如何导致下游泄露。在ConfAIde和PrivacyLens基准测试中，使用多个开源与闭源LLM进行实验，结果表明最优多智能体配置显著降低私密信息泄露（在GPT-4o上分别达18%和19%），同时保持公共内容完整性，优于单智能体基线。这些发现凸显了在多智能体系统中设计合理信息流对提升上下文隐私保护的潜力。
原文摘要 · Abstract (English)
Addressing contextual privacy concerns remains challenging in interactive settings where large language models (LLMs) process information from multiple sources (e.g., summarizing meetings with private and public information). We introduce a multi-agent framework that decomposes privacy reasoning into specialized subtasks (extraction, classification), reducing the information load on any single agent while enabling iterative validation and more reliable adherence to contextual privacy norms. To understand how privacy errors emerge and propagate, we conduct a systematic ablation over information-flow topologies, revealing when and why upstream detection mistakes cascade into downstream leakage. Experiments on the ConfAIde and PrivacyLens benchmark with several open-source and closed-sourced LLMs demonstrate that our best multi-agent configuration substantially reduces private information leakage (\textbf{18\%} on ConfAIde and \textbf{19\%} on PrivacyLens with GPT-4o) while preserving the fidelity of public content, outperforming single-agent baselines. These results highlight the promise of principled information-flow design in multi-agent systems for contextual privacy with LLMs.
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。