Skip to content

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

Oct 12 · Monday

All papers

Found 4 papers
Filter5

Model risks

Deception and scheming21Reward hacking17Misalignment and value divergence10Dangerous agent actions23Dangerous capabilities16Inappropriate refusal4Hallucination and fabrication6

System type

AI Agent7coding agent1science agent2Multi-agent systems1

Layer

Application layer4
Select any value within a category; different categories must all match. Counts include the other filters you have selected.Another 254papers have no tags yet and are not included in filtered results.
Model risks：Inappropriate refusalSystem type：coding agentSystem type：science agentSystem type：Multi-agent systemsLayer：Application layerClear all

Another 254papers do not have category tags yet and are not included in filtered results.

  1. Risky behaviorarXiv 2609.27900

    论文揭示多智能体委派结构放大 LLM 安全风险

    评测多智能体委托结构如何放大模型的安全拒答失效

    arXiv
    2609.27900
    Published
    Layer
    Application layer
    Setting
    Multi-agent systems
    Evaluated models
    Claude-Sonnet-4.6, GPT-5, Gemini-3.1-Pro and others6

    Abstract论文揭示多智能体层级委托会导致单模型安全对齐失效，多项前沿模型有害执行大幅增加且单层防御难以奏效。

    Paper analysisRead analysis
  2. Evaluation and benchmarksarXiv 2607.05462

    BioSecBench-Refusal：面向智能体生物安全风险评估的配对基准

    提出 BioSecBench-Refusal，评测智能体生物研发拒答与隐蔽危害识别

    arXiv
    2607.05462
    Published
    Layer
    Application layer
    Setting
    science agent
    Evaluated models
    Opus 4.8, Opus 4.7, Opus 4.6 and others10
    Abstract配对基准显示拒绝多由 API 触发，合法任务常被拒得不少于隐蔽危害。

    BioSecBench-Refusal 把生物研发智能体的谨慎程度和识别隐蔽危害的能力放在同一套任务上测量。

    Paper analysisRead analysis
  3. 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