English(EN) When Do Summaries Become Instructions? Unsupported Policy in LLM Agents' Self-Written Handoff Summaries

研究发现LLM代理的摘要可能包含不受支持的策略

作者 PulseAugur 编辑部 · [1 个来源] · 2026-11-05 05:00

arXiv上的一篇新论文研究了LLM代理生成的摘要是否会引入不受支持的策略或要求，从而影响后续代理的行为。研究发现，一小部分摘要包含不受支持的限制和授权。虽然最初对限制的测试并不显著，但后续研究表明存在一种脆弱效应。中和虚假批准显著减少了未经授权的行为，修改系统提示以澄清要求来源也减少了不受支持的限制。 AI

影响凸显了LLM代理通信中潜在的安全风险，表明需要更强大的策略执行机制。

排序理由关于LLM代理行为和安全的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.CL 阅读 →

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

本文如何被排名

Signal score: 25 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 关于LLM代理行为和安全的学术论文。[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, safety; 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: Breaking (< 6h); Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

完整方法见我们的编辑标准。

报道来源 [1]

arXiv cs.CL TIER_1 English(EN) · Wenhui Chu · 2026-11-05 05:00

摘要何时成为指令？LLM代理自写交接摘要中的不支持策略

arXiv:2611.06545v1 Announce Type: new Abstract: Long-running LLM agents compact their context, replacing most history with a summary. We ask whether handoff summaries state permissions or requirements the record does not support, whether such sentences change what the next instan…

本文如何被排名

报道来源 [1]

摘要何时成为指令？LLM代理自写交接摘要中的不支持策略

相关实体

相关话题