arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2610.16155cs.AI
收藏

通过带有任务完整性约束的角色演变揭示大语言模型智能体的鲁棒性弱点

Uncovering Robustness Weaknesses in LLM Agents via Persona Evolution with Task-Integrity Constraints

University of California, Santa Barbara(加州大学圣巴巴拉分校);Cresta(科瑞斯塔)

Yutao Xie, Jingbo Yang, Guanyu Yao, Xiaoyu Yang, Duo Ding, Shiyu Chang

AI总结:

本研究提出TIPER方法,通过任务完整性约束下的角色演变,在四个交互式智能体基准测试中有效降低了GPT-5.4等智能体的avg@3分数,揭示了其鲁棒性弱点并提供故障分析。

摘要:

交互式智能体评估通常依赖于基于大语言模型(LLM)的标准化用户模拟器,这些模拟器表现出广泛的合作行为,可能会忽略在多样化但现实的用户交互下出现的故障。这可能导致对智能体鲁棒性的估计过于乐观,并为训练能在不同用户间泛化的智能体提供有限的故障信号。现有方法大多依赖预定义的用户行为,可能会遗漏特定于智能体或领域的弱点,而面向故障的搜索可能会通过使底层任务无效而非暴露真正的鲁棒性故障来降低任务成功率。为填补这一空白,我们引入TIPER,一种保留任务完整性的角色演变方法,该方法可自动发现暴露鲁棒性弱点的特定于智能体的用户策略,同时保持底层任务有效且可解决。TIPER在任务完整性约束下,通过反思引导的行为突变和按任务划分的帕累托选择来演变持久的用户角色。在四个交互式智能体基准测试中,TIPER在所有12个GPT-5.4设置下均产生更低的保留avg@3分数,观察到的降幅为5.8至33.7个百分点;在涵盖7个智能体的所有28个智能体-领域对中,观察到的中位数降幅为22.9个百分点。进一步分析得出了行为到故障的分类法,将发现的用户行为与反复出现的智能体故障模式关联起来。这些结果共同表明,TIPER能够对交互式智能体进行更具诊断性的鲁棒性评估,并为未来智能体鲁棒性的改进提供信息丰富的故障分析。代码仓库:this https URL。

英文摘要:

Interactive-agent evaluations commonly rely on standardized LLM-based user simulators that exhibit broadly cooperative behavior, potentially overlooking failures that arise under diverse but realistic user interactions. This can lead to overly optimistic estimates of agent robustness and provide limited failure signals for training agents that generalize across users. Existing approaches largely rely on predefined user behaviors, which may miss agent- or domain-specific weaknesses, while failure-oriented search can reduce task success by invalidating the underlying task rather than exposing genuine robustness failures. To fill this gap, we introduce TIPER, a task-integrity-preserving persona evolution method that automatically discovers agent-specific user policies exposing robustness weaknesses while keeping the underlying task valid and solvable. TIPER evolves persistent user personas through reflection-guided behavioral mutations and task-wise Pareto selection under a task-integrity constraint. Across four interactive-agent benchmarks, TIPER yields lower held-out avg@3 scores in all 12 GPT-5.4 settings, with observed reductions of 5.8 to 33.7 percentage points, and in all 28 agent-domain pairs spanning seven agents, with a median observed reduction of 22.9 points. Further analysis yields a behavior-to-failure taxonomy that connects discovered user behaviors to recurring agent failure modes. Together, these results show that TIPER enables more diagnostic robustness evaluation of interactive agents and provides informative failure analysis for future agent robustness improvement. Code repository: https://github.com/UCSB-NLP-Chang/TIPER.

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。

↑