PulseAugur
中
实时 10:26:32
English(EN) Can Failed Trajectories Reveal Where Agents Go Wrong? Agentic Failure Attribution Without Step-Level Labels

新的SOAP框架可在无步级标签的情况下识别LLM Agent的故障

作者 PulseAugur 编辑部 · [1 个来源] ·

研究人员开发了一个名为SOAP（Spectral scOring with Attention-guided Propagation）的新框架，用于归因LLM Agent的故障。该方法仅使用失败轨迹，无需步级标签或成功示例，即可识别导致任务失败的关键步骤。SOAP构建了一个谱参考空间，并为每个步骤分配错误分数，然后利用注意力模式将下游错误证据传播到其可能的上游来源，实验结果优于现有基线。 AI

影响 这项研究通过提供一种更有效的方法来识别故障点，有望显著提高LLM Agent的调试效率和可靠性。

排序理由 该集群包含一篇学术论文，详细介绍了一种新的LLM Agent故障归因方法。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.LG 阅读 →

AI 生成摘要 · Google Gemini · 来自 1 个来源。 我们如何撰写摘要 →

新的SOAP框架可在无步级标签的情况下识别LLM Agent的故障

本文如何被排名

Signal score
11 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
该集群包含一篇学术论文，详细介绍了一种新的LLM Agent故障归因方法。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration
Single-source cluster
Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics
paper, other
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
Same-day
Cluster formed today. Ranking reflects the current source set at time of score.

完整方法见我们的编辑标准。

报道来源 [1]

  1. arXiv cs.LG TIER_1 English(EN) · Do Nguyen-Thanh, Junlin Fang, Bing Li, Zhen Fang, Sean Du ·

    失败的轨迹能否揭示代理的错误所在？无需步级标签的代理失败归因

    arXiv:2610.13105v1 Announce Type: new Abstract: Failure attribution for LLM agents aims to identify decisive steps responsible for task failure, which is critical for debugging and improving agentic systems. Existing methods rely on costly LLM prompting, step-level error annotati…

相关实体

相关话题