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 2605.14865cs.AIcs.CL
收藏

人工智能代理的总体评估与故障诊断

Holistic Evaluation and Failure Diagnosis of AI Agents

Deepchecks
Netta Madvil, Gilad Dym, Alon Mecilati, Edo Dekel, Jonatan Liberman, Rotem Brazilay, Liron Schliesser, Max Svidlo, Shai Nir, Orel Shalom, Yaron Friedman, David … 展开作者

Netta Madvil, Gilad Dym, Alon Mecilati, Edo Dekel, Jonatan Liberman, Rotem Brazilay, Liron Schliesser, Max Svidlo, Shai Nir, Orel Shalom, Yaron Friedman, David Connack, Amos Rimon, Philip Tannor, Shir Chorev

AI总结:

本文提出一种综合评估框架,结合顶层代理诊断与底层跨度评估,通过分解分析实现独立跨度评估,提升长结构轨迹中的故障定位与分类准确性。

摘要:

人工智能代理执行复杂的多步骤过程,但当前评估不足:结果指标仅报告成功或失败而无法解释原因,过程级方法难以将故障类型与长结构轨迹中的精确位置关联。本文提出一种综合代理评估框架,结合顶层代理级诊断与底层跨度级评估,将分析分解为独立的跨度级评估。这种分解可扩展至任意长度的轨迹,并为每个判断生成跨度级理由。在TRAIL基准测试中,该框架在GAIA和SWE-Bench上所有指标均取得最佳成绩,相对于最强的先前基线,在类别F1上提升达38%,在定位准确性上提升达3.5倍,在联合定位-分类准确性上提升达12.5倍。每类分析显示,该框架在更多错误类别上领先于其他评估器。值得注意的是,同一前沿模型在本框架中用于定位时,其定位准确性比作为整体轨迹的单一判断者高数倍,表明评估方法而非模型能力是瓶颈。

英文摘要:

AI agents execute complex multi-step processes, but current evaluation falls short: outcome metrics report success or failure without explaining why, and process-level approaches struggle to connect failure types to their precise locations within long, structured traces. We present a holistic agent evaluation framework that pairs top-down agent-level diagnosis with bottom-up span-level evaluation, decomposing analysis into independent per-span assessments. This decomposition scales to traces of arbitrary length and produces span-level rationales for each verdict. On the TRAIL benchmark, our framework achieves state-of-the-art results across all metrics on both GAIA and SWE-Bench, with relative gains over the strongest prior baselines of up to 38% on category F1, up to 3.5x on localization accuracy, and up to 12.5x on joint localization-categorization accuracy. Per-category analysis shows our framework leading in more error categories than any other evaluator. Notably, the same frontier model achieves several times higher localization accuracy when used inside our framework than as a monolithic judge over the full trace, showing that evaluation methodology, not model capability, is the bottleneck.

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。

↑