PulseAugur
中
实时 18:04:14
English(EN) After the Attacker Leaves: Benchmarking Carryover Attacks on LLM Agents

新基准揭示大型语言模型代理中持续存在的提示注入风险

研究人员开发了 CarryoverBench,这是一个旨在评估提示注入攻击在大型语言模型代理中持久性的新基准。这些“残留攻击”可以在初始恶意输入很久之后跨会话重新触发,利用代理的内存文件或计划任务。该基准包括 12 个个人助理场景和 101 个跨越即时、自主和潜在三种触发通道的攻击任务。对五个前沿模型的测试表明,植入的有效载荷可以在多个回合中重新触发,而不会显着衰减,并且运行时层控制可能是有效防御所必需的。 AI

影响 凸显了大型语言模型代理中的关键安全漏洞,可能影响其在持久性应用中的安全部署。

排序理由 该集群包含一篇研究论文,详细介绍了用于评估大型语言模型代理特定类型安全漏洞的新基准。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.AI 阅读 →

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

新基准揭示大型语言模型代理中持续存在的提示注入风险

本文如何被排名

Signal score
0 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
该集群包含一篇研究论文,详细介绍了用于评估大型语言模型代理特定类型安全漏洞的新基准。[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
17 days old
Aged out of breaking-news scoring windows; ranking reflects the durable signal from the full source set.

完整方法见我们的编辑标准。

报道来源 [1]

  1. arXiv cs.AI TIER_1 English(EN) · Zhiyue Xu, Xukai Wang, Zhiyuan Huang, Mingjie Zhan, Jian Zhao, Xu-Yao Zhang ·

    攻击者离开后:对LLM代理的残留攻击进行基准测试

    arXiv:2610.28425v1 Announce Type: cross Abstract: Modern LLM agents persist state across sessions via two channels: long-lived memory files silently prepended to the system prompt, and scheduled tasks waking the agent without user input. A one-time indirect prompt injection commi…