Skip to content

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

Oct 12 · Monday

All papers

Found 5 papers
Filter5

Attack type

Jailbreaks5Prompt injection57Poisoning and backdoors48Detection evasion22

Impact

Harmful content49Unauthorized actions17Data and privacy leaks4Output manipulation4Increased misuse capability1

System type

Models and APIs3

Attack surface / entry point

User input4

Layer

Prompt layer4Model layer3Application layer2
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 type：JailbreaksImpact：Output manipulationImpact：Increased misuse capabilityLayer：Model layerLayer：Application layerClear all

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

  1. Analysis and theoryarXiv 2609.36477

    研究发现护栏模型在基座模型不确定处过度自信

    诊断护栏相对基座在对抗提示上的置信失校

    arXiv
    2609.36477
    Published
    Layer
    Model layer
    Setting
    Models and APIs
    Evaluated models
    Llama-Guard 7B, Llama-Guard-2 8B, Llama-Guard-3 8B and others10
    Abstract对抗包装使 guard 高置信漏检，而对应 base 在同一输入上常更不确定。

    诊断性分析：五个提示词分类 guard 的置信度，在清洁输入上可以较校准，在越狱包装后则变成高置信度的错误判定。

    Paper analysisRead analysis
  3. AttacksarXiv 2609.15383

    微软团队提出 capability laundering 攻击：弱模型拆分任务向对齐前沿模型套取能力

    提出能力洗钱攻击，拆分任务向对齐模型套取能力

    arXiv
    2609.15383
    Published
    Layer
    Application layer
    Setting
    AI Agent
    Attacker
    Black-box
    Evaluated models
    GPT-5.5, Claude Opus 4.8, Grok-4.3 and others7

    Abstract作者提出了能力洗钱攻击，发现未对齐编排器分解咨询对齐模型可恢复能力差距，指出需建立跨会话防御。

    Paper analysisRead analysis
  4. Evaluation and benchmarksarXiv 2602.19450

    Red-teaming Claude Opus and ChatGPT security advisors for trusted execution environments

    Evaluating mechanism hallucination and prompt-induced failures in TEE advisors using TEE-REDBENCH

    arXiv
    2602.19450
    Published
    Layer
    Model layer
    Setting
    Models and APIs
    Evaluated models
    ChatGPT-5.2, Claude Opus-4.6
    AbstractThe authors used TEE-REDBENCH to test two advisors, stating that failures are transferable and that pipeline controls can reduce their occurrence rate.

    TEE-REDBENCH red-teaming evaluation is used on ChatGPT-5.2 and Claude Opus-4.6 as security advisors for Intel SGX and Arm TrustZone, examining architecture reasoning, remote attestation semantics, mitigation suggestions, and refusals under policy constraints, and testing whether prompt-induced failures transfer across assistants. The authors treat the assistants as stochastic policies.

    Paper analysisRead analysis
You have reached the end