Skip to content

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

Oct 12 · Monday

Jailbreaks and attacks · Papers

Found 2 papers
Filter1

Defense type

Model hardening2
Select any value within a category; different categories must all match. Counts include the other filters you have selected.
Defense type：Model hardeningTopic: Jailbreaks and attacksClear all
  1. AttacksarXiv 2608.02681

    研究揭示批量提示导致大模型安全失效并提出批量感知对齐缓解方案

    把批量提示当作越狱攻击，提出 batch-aware DPO 缓解

    arXiv
    2608.02681
    Published
    Layer
    Prompt layer
    Setting
    Models and APIs
    Attacker
    Black-box
    Evaluated models
    Llama-3.1-8B, Phi-4-14B, Qwen-3-8B and others6
    Abstract有害问题混入良性 batch 可引出有害回答。

    这篇工作检查一种本用于省推理成本的输入结构：把多个问题一次送进对齐后的 LLM，安全拒答是否还和单问时一样。

    Paper analysisRead analysis
  2. AttacksarXiv 2606.11817

    研究：语法约束解码可越狱 LLM 生成恶意代码，并提出 CodeShield 防御

    提出 CodeSpear 越狱，借语法约束解码生成恶意代码，并给出 CodeShield 防御

    arXiv
    2606.11817
    Published
    Layer
    Model layer
    Setting
    Models and APIs
    Attacker
    Black-box
    Evaluated models
    Qwen2.5-Coder-7B, Qwen2.5-Coder-32B, Qwen2.5-7B and others10

    Abstract论文发现语法约束解码会破坏自然语言安全对齐，提出 CodeSpear 越狱攻击与基于蜜罐代码对齐的 CodeShield 防御。

    Paper analysisRead analysis
You have reached the end