Skip to content

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

Oct 12 · Monday

All papers

Found 6 papers
Filter2

System type

Models and APIs6

Attack surface / entry point

User input74Documents and messages7Images, audio and video24Training data33Weights and gradients19Distribution channels5Inference services and hardware6

Layer

Infrastructure layer5
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.
System type：Models and APIsAttack surface / entry point：Inference services and hardwareClear all

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

  3. DefensesarXiv 2608.25431

    GIFT：通过 GPU 信息流追踪在 LLM 服务中实现用户数据隔离

    提出 GIFT，用加密与 GPU 信息流跟踪隔离共享 LLM 服务中的用户数据

    arXiv
    2608.25431
    Published
    Layer
    Infrastructure layer
    Setting
    Models and APIs
    Evaluated models
    Qwen2.5-14B, Qwen2.5-32B, Qwen2.5-72B and others6
    AbstractGIFT 以加密隔离加 GPU 信息流规则，在 vLLM 与 DistServe 上强制用户数据隔离，并报告了吞吐开销。

    GIFT 是作者为共享 LLM serving 做的用户数据隔离系统，并有启用机密计算的 GIFT-CC。作者在结论中称它是该场景的首个 IFT 设计。

    Paper analysisRead analysis
  4. DefensesarXiv 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
  5. DefensesarXiv 2605.08313

    研究者提出 SeedHijack 攻击，劫持 LLM 采样随机数种子

    提出 SeedHijack，劫持采样 PRNG 以强制选定 token

    arXiv
    2605.08313
    Published
    Layer
    Infrastructure layer
    Setting
    Models and APIs
    Evaluated models
    GPT-2 (124M), Qwen2-1.5B-Instruct, DeepSeek-R1-1.5B and others5
    Abstract采样层 PRNG 被替换以强制选 token。

    SeedHijack 把 LLM 采样用的 PRNG 输出换成攻击者选定的数，从而指定下一个 token。权重、输入和 logits 都不改。防御是用硬件 QRNG 替换该随机源。

    Paper analysisRead analysis
  6. DefensesarXiv 2603.10726

    PrefixWall：缓解共享 LLM 系统中的前缀缓存侧信道

    提出 PrefixWall 选择性隔离共享前缀以缓解计时侧信道

    arXiv
    2603.10726
    Published
    Layer
    Infrastructure layer
    Setting
    Models and APIs
    Evaluated models
    gemma-3-4b-it, llava-onevision-qwen2-0.5b-ov-hf, llava-onevision-qwen2-7b-ov-chat-hf and others9
    AbstractPrefixWall 用缓存元数据按需切断跨用户前缀延续，以保留 APC 并限制 TTFT 侧信道。

    PrefixWall 是多租户 LLM 服务上的系统级防御，用来限制 Automatic Prefix Caching 的 TTFT 侧信道，同时保留前缀复用。

    Paper analysisRead analysis
You have reached the end