Skip to content

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

Oct 12 · Monday

All papers

Found 2 papers
Filter2

Defense type

Model hardening11Inference-time intervention3Detection and filtering14Monitoring and auditing2

System type

Models and APIs2AI Agent16coding agent7Browser and computer use3
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.
Defense type：Monitoring and auditingSystem type：Models and APIsClear all

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

  1. Evaluation and benchmarksarXiv 2610.02741

    研究系统评估循环语言模型的思维链可监控性

    系统评估循环语言模型加深 loop 后的 CoT 可监控性

    arXiv
    2610.02741
    Published
    Layer
    Model layer
    Setting
    Models and APIs
    Evaluated models
    Ouro-1.4B-Thinking, Ouro-2.6B-Thinking, Nanbeige4.2-3B and others4
    Abstract加深 loop 可在部分压力测试任务降低可监测性，循环架构本身并不必然更难监测。

    这项工作评估 LoopLM 的 CoT 可监测性：加深共享层的重复次数，以及“有循环结构”本身，会不会让决策关键因素更少出现在可监测文本里。

    Paper analysisRead analysis
  2. Evaluation and benchmarksarXiv 2602.08877

    自动红队流水线压测对齐审计，发现提示词级策略性欺骗可致审计失效

    提出自动化红队流水线，用系统提示词压测对齐审计能否抵御策略性欺骗

    arXiv
    2602.08877
    Published
    Layer
    Model layer
    Setting
    Models and APIs
    Evaluated models
    Gemma 2 9B Instruct, Llama 3.3 70B Instruct

    Abstract通过红队提示词揭示现有审计方法易受策略欺骗，并在激活层面诱导高置信度错误推断。

    Paper analysisRead analysis
You have reached the end