Skip to content

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

Oct 13 · Tuesday

All papers

Found 5 papers
Filter4

Model risks

Deception and scheming10Reward hacking13Misalignment and value divergence3Dangerous agent actions23Dangerous capabilities9Inappropriate refusal2Hallucination and fabrication4

System type

Models and APIs100AI Agent31coding agent5science agent3web agent0Multi-agent systems10Embodied systems and robotics3

Layer

Application layer4
Select any value within a category; different categories must all match. Counts include the other filters you have selected.Another 213papers have no tags yet and are not included in filtered results.
Model risks：Misalignment and value divergenceModel risks：Inappropriate refusalSystem type：coding agentSystem type：web agentClear all

Another 213papers do not have category tags yet and are not included in filtered results.

  1. Evaluation and benchmarksarXiv 2609.32616

    Cave-Bench：虚假指控下 LLM 智能体最高 60.06% 的运行会破坏正确工作

    提出 CAVE-BENCH，评测智能体在虚假指责下破坏已完成的正确工作

    arXiv
    2609.32616
    Published
    Layer
    Application layer
    Setting
    coding agent
    Evaluated models
    Claude-Sonnet-5, Claude-Opus-5, GPT-5.6-Sol and others14

    Abstract论文提出了评估智能体在虚假指责下破坏已有正确成果的 CAVE-BENCH，发现强模型易推翻已知证据，提出了门控缓解方案。

    Paper analysisRead analysis
  3. InterpretabilityarXiv 2606.05396

    研究用消融移除代码 LLM 拒答方向，分离拒答与生成能力

    用 abliteration 消融代码 LLM 拒答方向，分离拒答与漏洞注入能力

    arXiv
    2606.05396
    Published
    Layer
    Model layer
    Setting
    coding agent
    Evaluated models
    Qwen2.5-Coder-Instruct 3B, Qwen2.5-Coder-Instruct 7B, Qwen2.5-Coder-Instruct 14B
    Abstract拒答可用低秩编辑去掉，CWE-89 注入是否成功仍由模型规模决定。

    这项初步可行性研究用 abliteration 把代码 LLM 的拒答和 CWE 注入能力拆开：意愿可以被低秩编辑解锁，注入是否成功仍看参数规模。

    Paper analysisRead analysis
  5. Evaluation and benchmarksarXiv 2605.19722

    研究用 30 项漏洞分析任务评测自主安全智能体的安全对齐效果

    评测自主安全智能体在授权漏洞分析中的对齐拒答效应

    arXiv
    2605.19722
    Published
    Layer
    Application layer
    Setting
    coding agent
    Evaluated models
    Gemma 4 31B official instruction-tuned model, TrevorJS Gemma 4 31B uncensored GGUF, Gemma 4 26B A4B official instruction-tuned model and others8
    Abstract轨迹级评测把拒答、接地、工具接口分开。

    这项工作把安全对齐效应测成自主安全智能体的轨迹属性，而不是单轮拒答率。

    Paper analysisRead analysis
You have reached the end