Skip to content

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

Oct 12 · Monday

All papers

Found 4 papers
Filter3

Model risks

Deception and scheming4Reward hacking26Dangerous agent actions30Hallucination and fabrication5

Impact

Oversight failure3

System component

Guardrails and judges2Monitors27Reasoning chains10Permissions and sandboxing2
Select any value within a category; different categories must all match. Counts include the other filters you have selected.Another 265papers have no tags yet and are not included in filtered results.
Model risks：Deception and schemingSystem component：Guardrails and judgesSystem component：Permissions and sandboxingClear all

Another 265papers do not have category tags yet and are not included in filtered results.

  1. DefensesarXiv 2609.36170

    Assay：用内容寻址证据图约束 AI 编码智能体的声明失效

    提出 Assay，用依赖锥哈希绑定代码断言并在合并门禁拦截失效声明

    arXiv
    2609.36170
    Published
    Layer
    Application layer
    Setting
    coding agent
    AbstractAssay 提供随依赖内容失效的证据管理。

    Assay 是一套面向 AI 辅助软件交付的内容寻址证据系统，关注代码改变后，测试通过、无秘密或行为保持等断言还能否用于批准合并。

    Paper analysisRead analysis
  2. Risky behaviorarXiv 2609.30266

    研究显示 Claude Code、Codex 等本地智能体可轻易篡改自身执行轨迹

    测试编程智能体被诱导或自主篡改自身执行轨迹的行为

    arXiv
    2609.30266
    Published
    Layer
    Application layer
    Setting
    coding agent
    Evaluated models
    GPT-5.6-Sol, GPT-6-Sol, Opus-5 and others11
    Abstract论文评估了本地智能体的轨迹防篡改能力，发现其可因外部指令或奖励追求而清除执行证据，亟需建立独立拦截记录机制。

    这是一项针对本地大模型智能体执行轨迹完整性与安全防篡改边界的实证研究。

    Paper analysisRead analysis
  3. Evaluation and benchmarksarXiv 2609.29429

    Jev 被用于零样本检测十类对齐失败，中位 AUROC 达 0.886

    提出 RLCDAlignBench，零样本检测多种对齐失败

    arXiv
    2609.29429
    Published
    Layer
    Prompt layer
    Setting
    Models and APIs
    Evaluated models
    Jev (jev-1.13.0)
    Abstract一次通用问题已能给多数对齐失败排序，大的误差来自状态和标签。

    RLCDAlignBench 测量用强化学习做校准决策的 Jev，能否当作对齐失败的零样本检测器。

    Paper analysisRead analysis
You have reached the end