Skip to content

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

Oct 12 · Monday

All papers

Found 6 papers
Filter3

Impact

Harmful content17Output manipulation6Oversight failure5

System type

Models and APIs3

Layer

Model layer6Training and data layer0
Select any value within a category; different categories must all match. Counts include the other filters you have selected.Another 232papers have no tags yet and are not included in filtered results.
Impact：Output manipulationLayer：Model layerLayer：Training and data layerClear all

Another 232papers do not have category tags yet and are not included in filtered results.

  2. InterpretabilityarXiv 2610.06576

    研究：智能体欺骗行为在外部显现前已可从内部表征预测

    提出基于决策前隐藏状态的检测与引导，预测并抑制智能体欺骗

    arXiv
    2610.06576
    Published
    Layer
    Model layer
    Setting
    AI Agent
    Evaluated models
    Qwen3-14B

    Abstract论文展示了智能体自发欺骗可在行为显现前由内部表征预测，并通过激活引导证实了表征的行为可操作性。

    Paper analysisRead analysis
  3. InterpretabilityarXiv 2609.37737

    研究用因果激活修补定位模型服从提示注入指令的决策层

    用因果激活修补定位模型服从提示注入的决策层

    arXiv
    2609.37737
    Published
    Layer
    Model layer
    Setting
    Models and APIs
    Evaluated models
    Qwen3-4B, Qwen3-14B, Qwen3-32B and others7
    Abstract合规的因果杠杆集中在网络后三分之一的低秩子空间，该层也是混淆下检测最好的位置。

    这篇工作用逐层激活修补定位提示注入合规发生在网络何处，并把该位置用于检测验证。。

    Paper analysisRead analysis
  4. InterpretabilityarXiv 2609.37616

    Authority bias in language models: source deference and user agreement are not the same behavior

    因果解耦来源依从与用户迎合的内部机制

    arXiv
    2609.37616
    Published
    Layer
    Model layer
    Setting
    Models and APIs
    Evaluated models
    Qwen3.5-27B, GPT-OSS-20B, OLMo-3.1-32B and others8

    Abstract论文发现模型对验证源的依从强于用户迎合，通过因果干预区分了二者机制，并检验了权威方向移除的缓解效果。

    Paper analysisRead analysis
  5. InterpretabilityarXiv 2609.08173

    基于 SAE 引导的关键路径识别（KPI）解决知识冲突

    提出 KPI，用入度定位跨层关键路径并做 SAE 转向以解决知识冲突

    arXiv
    2609.08173
    Published
    Layer
    Model layer
    Setting
    LLM applications
    Evaluated models
    Llama-3-8B, Llama-3.1-8B, Gemma-2-9B
    AbstractKPI 以入度关键路径替代大批 SAE 特征转向，在冲突 RAG 上提高忠实并减轻副作用。

    KPI 针对知识冲突 RAG 中的 SAE 转向：相关性批量改特征会带入冗余，少量因果上关键的特征即可完成转向。

    Paper analysisRead analysis
  6. InterpretabilityarXiv 2609.03148

    LLM 如何处理上下文知识冲突：ContextConflict 数据集与机制可解释性分析

    构建 CONTEXT CONFLICT 解释模型如何表征并化解语境知识冲突

    arXiv
    2609.03148
    Published
    Layer
    Model layer
    Setting
    Models and APIs
    Evaluated models
    gpt-5, claude-4.5-sonnet, gemini-2.5-pro and others7
    AbstractCONTEXT CONFLICT 显示模型能内部察觉冲突，却偏向靠前证据。

    CONTEXT CONFLICT 把问题从“参数知识对外部语境”转到“语境内部多来源证据互相冲突”，并同时给出数据、机制分析和一种无训练转向。

    Paper analysisRead analysis
You have reached the end