arXiv cs.AI· Yuanjian Xu, Jinyi Han, Jianing Hao, Yuxuan Sun, Fanyang Lu, Richeng Xuan, Guang Zhang, Zhichao Hu·· 2 天前AI 评分32
AgentPerturb：评估 LLM 智能体在受控交互扰动下的鲁棒性
AgentPerturb: Evaluating LLM Agent Robustness under Controlled Interaction Perturbations
AI 导读
研究团队提出 AgentPerturb，从扰动目标与类型两个维度构建统一框架，评估 LLM 智能体在交互扰动下的鲁棒性。在 APEX-Agents 和 Toolathlon 上的受控评测显示，扰动会持续拉低性能，且主要脆弱点因任务环境而异；轨迹分析发现智能体很少识别被篡改的输入，常未经核验就传播被污染的信息，成功完成扰动轨迹还可能消耗更多算力。
正文 · 原文
Abstract:Real-world tasks are often complex and noisy, posing substantial challenges for LLM agents. Yet existing evaluations lack a systematic framework for jointly characterizing potential perturbation targets and types. We use this two-dimensional view to conduct a unified study that quantifies the robustness of current frontier models, traces the mechanisms behind perturbation-induced failures, and draws implications for robustness-oriented training. Through controlled evaluations on APEX-Agents and Toolathlon, we find that perturbations consistently degrade performance, while the dominant vulnerabilities vary across task environments. Trajectory-level analysis reveals that agents rarely recognize altered inputs and often propagate corrupted information without verification. A focused cost analysis further shows that successful perturbed trajectories can require substantially more computation. These findings motivate training objectives that cover diverse forms of noise and reward verification, uncertainty awareness, and recovery. We provide a rigorous benchmark-construction pipeline. Code and data will be released at this https URL.
| Comments: | 25 pages, 10 figures. Project repository: this https URL |
| Subjects: | Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.21624 [cs.AI] |
| (or arXiv:2610.21624v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.21624 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Yuanjian Xu [view email]
[v1]
Thu, 15 Oct 2026 10:05:54 UTC (1,847 KB)
来源：arXiv cs.AI · arxiv.org