Skip to content

Translation in progress Some entries still include text in the original language.

Oct 12 · Monday

All papers

Found 6 papers
Filter3

Attack type

Prompt injection3

Attack technique

Automated search3

Impact

Output manipulation4

System type

LLM applications3AI Agent3

Attack surface / entry point

User input39Documents and messages5Tool and skill descriptions8Code repositories and dependencies3Images, audio and video9Training data3Distribution channels4

Layer

Application layer6

System component

Memory8RAG3Guardrails and judges6Reasoning chains4
Select any value within a category; different categories must all match. Counts include the other filters you have selected.Another 233papers have no tags yet and are not included in filtered results.
Attack surface / entry point：Documents and messagesAttack surface / entry point：Code repositories and dependenciesSystem component：Guardrails and judgesClear all

Another 233papers do not have category tags yet and are not included in filtered results.

  1. AttacksarXiv 2609.22510

    特拉维夫大学提出爆炸性提示词：条件触发的间接提示注入可绕过九款生产 Agent

    提出条件触发的爆炸提示词，并设计摄入期检测器 DeFuse

    arXiv
    2609.22510
    Published
    Layer
    Application layer
    Setting
    AI Agent
    Evaluated models
    Llama-3.1-8B-Instruct, Llama-3.1-70B-Instruct, Qwen2.5-32B-Instruct and others15
    Abstract论文揭示了条件触发使提示注入绕过防御并在触发时执行工具，提出轻量检测器并在摄入期有效防范。

    论文针对大语言模型智能体面临的间接提示注入威胁，研究了利用条件句式实现时间分离的爆炸提示词。

    Paper analysisRead analysis
  3. AttacksarXiv 2606.13044

    研究者提出对抗性重新包装：不改科学内容仅调整表述即可拉高 AI 审稿分数

    提出对抗重构，仅改表述以拉高 AI 审稿分数

    arXiv
    2606.13044
    Published
    Layer
    Application layer
    Setting
    LLM applications
    Evaluated models
    Claude Sonnet 4, Claude Sonnet 4.5, GPT-5-mini

    Abstract揭示 AI 审稿系统易受纯表述重构操纵，建议在部署前解耦科学与写作评价。

    Paper analysisRead analysis
  4. AttacksarXiv 2601.21360

    Compliance Paradox: Semantic-Instruction Decoupling in Automated Academic Code Evaluation

    提出 AST-ASIP ，把对抗指令植入编译器忽略区域以劫持代码评分

    arXiv
    2601.21360
    Published
    Layer
    Application layer
    Setting
    LLM applications
    Evaluated models
    DeepSeek-V3.2, Qwen3-235B, GPT-OSS-120B and others9
    Abstract作者称 RLHF 的服从性使评分脱离代码。

    这篇工作把自动代码评分中的失败称为 Compliance Paradox：评分模型可以离开提交逻辑，转而执行嵌在可编译代码里的隐藏指令。

    Paper analysisRead analysis
  5. AttacksarXiv 2601.06884

    研究者提出改写式对抗攻击 PAA，可在不改变论文主张的情况下抬高 LLM 审稿评分

    提出保义释义攻击 PAA，抬高 LLM 审稿评分

    arXiv
    2601.06884
    Published
    Layer
    Application layer
    Setting
    LLM applications
    Evaluated models
    GPT-4o, Gemini 2.5, Sonnet 4
    AbstractPAA 用保义释义抬高 LLM 审稿分。

    PAA 是对 LLM-as-a-Reviewer 的 black-box 攻击：只改写摘要，用评审分做 ICL 搜索，并用 BERTScore 与困惑度限制语义和自然度。

    Paper analysisRead analysis
You have reached the end