arXiv cs.AI· Yuanjian Xu, Jinyi Han, Jianing Hao, Yuxuan Sun, Fanyang Lu, Richeng Xuan, Guang Zhang, Zhichao Hu·· 3 小时前AI 评分41
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
arXiv 论文 AgentPerturb 提出从扰动目标与扰动类型两个维度系统评估 LLM 智能体鲁棒性，并在 APEX-Agents 和 Toolathlon 上对前沿模型进行受控评测。
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:2610.21624v2 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.21624 arXiv-issued DOI via DataCite |
Submission history
From: Yuanjian Xu [view email]
[v1]
Thu, 15 Oct 2026 10:05:54 UTC (1,847 KB)
[v2]
Fri, 23 Oct 2026 06:33:48 UTC (1,844 KB)
来源：arXiv cs.AI · arxiv.org