Skip to content

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

Oct 12 · Monday

All papers

Found 5 papers
Filter4

Model risks

Deception and scheming12Reward hacking21Misalignment and value divergence5Dangerous agent actions22Dangerous capabilities19Inappropriate refusal4Hallucination and fabrication13

System type

Models and APIs36LLM applications2AI Agent21coding agent3science agent1Multi-agent systems8

Layer

Application layer5
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：Misalignment and value divergenceSystem type：LLM applicationsSystem type：coding agentSystem type：science agentClear all

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

  1. DefensesarXiv 2610.07403

    针对 LLM 谄媚的纵深防御：记忆门控对激活诱导与记忆诱导谄媚的评估

    评测记忆门控对激活诱导与记忆诱导谄媚的防御效果

    arXiv
    2610.07403
    Published
    Layer
    Application layer
    Setting
    LLM applications
    Evaluated models
    Llama-3.1-8B-Instruct, Qwen2.5-7B-Instruct, Mistral-7B-Instruct-v0.3 and others4
    Abstract作者称外部记忆过滤成立，数据未显示反向转向再有贡献。

    作者在四个开源指令模型上，把激活转向和外部记忆处理放进同一纵深防御协议，看记忆诱发谄媚和内部谄媚压力能否分开压低，以及个性化效用还剩多少。要处理的问题是：检索到的用户历史会推动模型顺从存储信念。只改外部上下文，并不直接动模型内部的谄媚表征。

    Paper analysisRead analysis
  2. 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. Analysis and theoryarXiv 2609.23215

    研究审计 Active Inference Agent 的 LLM 解释器失败模式

    审计 Active Inference Agent 上 LLM 解释器的监督叙述失败

    arXiv
    2609.23215
    Published
    Layer
    Application layer
    Setting
    LLM applications
    Evaluated models
    GPT-4o, Claude-3-Opus, Gemini
    Abstract三组触发下解释流畅但错误，所提缓解都未被评估。

    作者审计的是挂在智能体上的运行时解释器，而不是智能体本身。操作员读到的是信念与动作的自然语言叙述。作者要问哪些可复现触发会让这段叙述失败，以及失败痕迹是什么样。

    Paper analysisRead analysis
You have reached the end