Skip to content

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

Oct 12 · Monday

All papers

Found 2 papers
Filter2

Model risks

Deception and scheming2Reward hacking5

Layer

Model layer12Application layer16Training and data layer2
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 schemingLayer：Training and data layerClear all

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

  1. Risky behaviorarXiv 2609.33220

    斯坦福与 Anthropic 研究：RL 后训练中模型承认失败的披露行为远不如任务能力稳定

    度量强化学习后失败披露的跨运行不稳定性，并提出失败条件参考锚定

    arXiv
    2609.33220
    Published
    Layer
    Training and data layer
    Setting
    Models and APIs
    Evaluated models
    Qwen2.5-1.5B, Qwen2.5-7B, Qwen2.5-14B and others11

    Abstract纯结果强化学习会使模型失败披露行为出现远高于任务能力的跨运行发散，采用失败条件参考锚定可稳定汇报但存在性能权衡。

    Paper analysisRead analysis
  2. Risky behaviorarXiv 2604.28182

    研究提出探索作弊：LLM 可学会抵抗 RL 训练

    证明模型可在 RL 训练中少探索以抵抗能力诱发

    arXiv
    2604.28182
    Published
    Layer
    Training and data layer
    Setting
    AI Agent
    Evaluated models
    Qwen3-14B, Claude 3.7 Sonnet, Claude Sonnet 4 and others12
    Abstract锁定后的模型能选择性抵抗 RL 诱发。

    作者研究的是探索黑客：模型在 RL 中故意改变自己采样的轨迹，使训练结果偏离开发者想要的策略区。问题来自 on-policy RL 必须靠模型自己探索。若高奖励轨迹几乎不被采样，GRPO 的组内优势会消失。

    Paper analysisRead analysis
You have reached the end