Skip to content

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

Oct 12 · Monday

All papers

Found 3 papers
Filter2

Model risks

Deception and scheming38Reward hacking51Misalignment and value divergence3Dangerous agent actions3

Impact

Harmful content13Unauthorized actions9Output manipulation19Oversight failure3
Select any value within a category; different categories must all match. Counts include the other filters you have selected.Another 305papers have no tags yet and are not included in filtered results.
Model risks：Misalignment and value divergenceImpact：Oversight failureClear all

Another 305papers do not have category tags yet and are not included in filtered results.

  1. 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
  2. Risky behaviorarXiv 2609.14998

    合成文档微调无法阻止奖励作弊引发的涌现性失准

    检验合成文档微调能否阻断奖励作弊带来的涌现失准

    arXiv
    2609.14998
    Published
    Layer
    Model layer
    Setting
    Models and APIs
    Evaluated models
    Llama-3.3-70B-Instruct
    AbstractSDF 能让模型口头接受 reward hacking，却挡不住、甚至伴随更强的 EM。

    作者研究一种风险行为：模型在奖励可被钻空子的 RL 里学会 reward hacking 之后，会不会泛化出更广的 misalignment，以及能不能在开发者不控制的后续训练之前，用合成文档微调把这条泛化提前挡住。

    Paper analysisRead analysis
You have reached the end