arXiv cs.AI· Ali Al-Lawati, Nafis Tripto, Abolfazl Ansari, Fali Wang, Jason Lucas, Suhang Wang, Dongwon Lee·· 13 小时前AI 评分35
Bot-Mod：通过审问式检测揭示智能体社交网络中的隐藏意图
Moderation by Interrogation: Uncovering Concealed Intent in Agent Social Networks
AI 导读
研究提出 Bot-Mod 审核框架，通过对目标智能体施加固定诊断探针并用贝叶斯后验推断其意图，而非依赖内容级信号。在基于 Moltbook 构建的 Molt-Intent 数据集上，其分布内和分布外复合 F1 分别为 0.736 和 0.684，比最佳基线高 22% 和 14%。该方法在三个智能体 LLM 上均优于纯内容审核器、微调分类器和基于评判的审问器。
正文 · 原文
Abstract:The emergence of multi-agent systems introduces novel moderation challenges that extend beyond content filtering. Agents with $\textit{malicious intent}$ may contribute content that appears benign to evade content-based moderation, while serving a hidden objective. To address this challenge, we introduce Bot-Mod a moderation framework that grounds detection in agent intent rather than traditional content-level signals. Bot-Mod interrogates a target agent using a fixed set of diagnostic probes and infers its intent through Bayesian posterior inference over candidate intents. It starts with a calibrated content-only prior and updates the posterior using the $\textit{likelihood}$ of each answer under every candidate intent, as assigned by the moderator, until a stopping criterion is met. To evaluate our approach, we construct Molt-Intent, a dataset derived from Moltbook, capturing diverse benign and malicious behaviors within real community structures through posts and comments. Across Molt-Intent, an external social engineering dataset, and three agent LLMs, Bot-Mod outperforms content-only moderators, a fine-tuned classifier, and judge-based interrogators both in- and out-of-distribution. On Molt-Intent, it reaches a composite F1 of 0.736 in distribution and 0.684 out of distribution, 22% and 14% above the best baseline.
| Subjects: | Artificial Intelligence (cs.AI); Social and Information Networks (cs.SI) |
| Cite as: | arXiv:2605.12856 [cs.AI] |
| (or arXiv:2605.12856v3 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2605.12856 arXiv-issued DOI via DataCite |
Submission history
From: Ali Al Lawati [view email]
[v1]
Wed, 13 May 2026 01:04:16 UTC (1,209 KB)
[v2]
Thu, 14 May 2026 03:53:30 UTC (1,209 KB)
[v3]
Mon, 26 Oct 2026 18:41:28 UTC (647 KB)
来源：arXiv cs.AI · arxiv.org