Skip to content

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

Oct 12 · Monday

All papers

Found 1 paper
Filter5

Model risks

Deception and scheming19Reward hacking13Misalignment and value divergence10Dangerous agent actions18Inappropriate refusal1Hallucination and fabrication4

Layer

Application layer0Training and data layer1

System component

Memory0Monitors1Vision4Multimodal4
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：Inappropriate refusalLayer：Application layerLayer：Training and data layerSystem component：MemorySystem component：MonitorsClear all

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

  1. DefensesarXiv 2609.38645

    用探针引导微调对齐模型且不损失可监控性

    用持续更新的探针做微调，降低有害性并保持线性可监测性

    arXiv
    2609.38645
    Published
    Layer
    Training and data layer
    Setting
    Models and APIs
    Evaluated models
    Mistral 7B Instruct v0.1, Llama 3 8B Instruct Abliterated, Qwen3-14B
    Abstract持续重拟合的探针可降低有害性、提高诚实性，训练后概念仍能被线性探针读出。

    Probe-guided fine-tuning 把探针分数当作唯一对齐信号，用来推动 harmlessness 与 honesty，并检查训练后这些概念是否仍线性可读。

    Paper analysisRead analysis
You have reached the end