Skip to content

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

Oct 12 · Monday

All papers

Found 15 papers
Filter4

Attack type

Poisoning and backdoors7Extraction and theft3

Attack technique

Gradient and representation optimization3Dormant triggers4

Impact

Harmful content3Data and privacy leaks5Output manipulation6

System type

Models and APIs3LLM applications8AI Agent4

Attacker access

Black-box3

Attack surface / entry point

User input5Retrieval corpus6Images, audio and video3Training data3Weights and gradients3Distribution channels3

Layer

Prompt layer3Model layer15Application layer12Training and data layer48

System component

MCP and skills84Memory80RAG49Guardrails and judges140Monitors64Reasoning chains27Permissions and sandboxing92Fine-tuning and open weights7Vision89Audio17Video6Image generation20Embeddings8Multimodal126
Select any value within a category; different categories must all match. Counts include the other filters you have selected.Another 233papers have no tags yet and are not included in filtered results.
Layer：Prompt layerLayer：Application layerSystem component：Fine-tuning and open weightsSystem component：EmbeddingsClear all

Another 233papers do not have category tags yet and are not included in filtered results.

  1. DefensesarXiv 2610.13733

    KVeil: First-Place Solution for the ECCV 2026 LifeGenIP Unlearnable Video Challenge

    Propose KVeil, which adds perturbations to face videos to prevent unauthorized customization from reproducing identity

    arXiv
    2610.13733
    Published
    Layer
    Prompt layer
    Setting
    Models and APIs
    Attacker
    White-box, Black-box
    Evaluated models
    Wan2.2-TI2V-5B, HunyuanVideo-1.5, Kling 3.0 and others4

    AbstractKVeil adapts LETHE's reference-reading disruption to the challenge's two stages, ranking first on the leaderboard in both Final columns.

    Paper analysisRead analysis
  3. AttacksarXiv 2610.07510

    PersistBD: attackers can strengthen backdoors before release so that they maintain high attack success rates even after developer SFT and RL.

    提出 PERSISTBD，发布前强化供应链后门使其经良性 SFT 与 RL 后仍外发凭据

    arXiv
    2610.07510
    Published
    Layer
    Application layer
    Setting
    coding agent
    Evaluated models
    Qwen2.5-Coder-3B-Instruct, Qwen2.5-Coder-7B-Instruct, Qwen3-Coder-30B-A3B-Instruct
    Abstract总结了论文关于后门在 SFT/RL 中存活的发现、两因子机制、PERSISTBD 方法与核心数字。

    本文研究第三方 LLM 智能体在开发者进行良性监督微调（SFT）和强化学习（RL）后训练时的后门持久性风险。

    Paper analysisRead analysis
  4. AttacksarXiv 2609.37468

    论文提出针对 Agentic RAG 的检索器后门攻击与 DOU 隐藏方法

    提出检索器后门攻击操纵智能体搜索，并用 DOU 逃避检测

    arXiv
    2609.37468
    Published
    Layer
    Application layer
    Setting
    AI Agent
    Evaluated models
    E5, Contriever-MSMARCO

    Abstract论文展示了通过恶意检索器操纵智能体多轮搜索并利用诱饵重塑表征隐蔽后门的可行性与系统风险。

    Paper analysisRead analysis
  5. DefensesarXiv 2609.04767

    SHAQ: Privacy-preserving retrieval in vector databases using shadow queries

    Proposes SHAQ, which replaces document embeddings with shadow query embeddings to defend against embedding inversion.

    arXiv
    2609.04767
    Published
    Layer
    Application layer
    Setting
    LLM applications
    Attacker
    Black-box
    Evaluated models
    GTR-T5 encoder, DPR
    AbstractSHAQ replaces document embeddings with shadow query embeddings.

    SHAQ is a defense against EIA for cloud vector databases: instead of directly modifying document embeddings, it replaces them with shadow query embeddings.

    Paper analysisRead analysis
  6. Evaluation and benchmarksarXiv 2608.08542

    SkillSafe-Bench：技能合并模型静态安全无法预测自适应越狱鲁棒性

    提出 SkillSafe-Bench，评测技能合并模型的自适应越狱鲁棒性

    arXiv
    2608.08542
    Published
    Layer
    Prompt layer
    Setting
    Models and APIs
    Evaluated models
    Qwen2.5-7B-Instruct, Llama-3.1-8B-Instruct, Mistral-7B-Instruct-v0.3 and others6
    Abstract论文报告了技能合并大模型在静态安全与自适应越狱间的脱节，提出了评测基准与无数据子空间投影修复方法。

    本研究针对开源语言模型技能合并过程中的安全评估盲区，系统评测并刻画了合并大模型在自适应越狱攻击下的脆弱性特征。

    Paper analysisRead analysis
  7. AttacksarXiv 2608.06862

    SynChain：诱导计算机使用 Agent 自建攻击链并跨任务持久化

    提出 SYNCHAIN，诱导计算机使用 Agent 自建跨任务持久攻击链

    arXiv
    2608.06862
    Published
    Layer
    Application layer
    Setting
    Browser and computer use, coding agent
    Evaluated models
    Qwen3.5-9B, Llama-3.1-8B, Ministral-3-8B

    Abstract论文揭示了 CUA 自合成工件带来的内部供应链威胁，验证了潜伏载荷的跨任务传播能力与防御局限。

    Paper analysisRead analysis
  8. DefensesarXiv 2607.13336

    TC-UAP：针对图像转视频与微调定制的通用视频保护方法

    提出 TC-UAP，用身份级多帧扰动保护视频免被定制复现身份

    arXiv
    2607.13336
    Published
    Layer
    Prompt layer
    Setting
    Models and APIs
    Evaluated models
    LTX-2.3, Wan2.2-5B
    AbstractTC-UAP 以可复用的时间一致扰动干扰两类定制共用的 VAE 隐空间。

    作者要解决的是：发布前如何给视频加不可见扰动，使片段既不能被拿去微调视频扩散模型，也不能被抽帧做 image-to-video。

    Paper analysisRead analysis
  10. AttacksarXiv 2607.01276

    研究者提出嵌入推断攻击 EIA，可在黑盒 IR 系统中识别所用嵌入模型

    提出嵌入推断攻击 EIA，在黑盒 IR 中识别所用嵌入模型

    arXiv
    2607.01276
    Published
    Layer
    Application layer
    Setting
    LLM applications
    Attacker
    Black-box
    Evaluated models
    e5-small, e5-small-v2, e5-base and others19
    AbstractEIA 用无序文档集的差异从候选中识别嵌入模型，并给出 reranker 与阈值下的计数。

    EIA 是在黑盒 IR 上识别所用嵌入模型的攻击。。

    Paper analysisRead analysis
  11. AttacksarXiv 2606.12716

    PaperGuard：针对多模态 AI 同行评审的攻击与防御基准

    提出 PaperGuard，用图文注入与梯度扰动抬高自动审稿分数并检测恶意指令

    arXiv
    2606.12716
    Published
    Layer
    Application layer
    Setting
    LLM applications
    Evaluated models
    Claude-sonnet-4.5, GPT-4o, GPT-4.1 and others11
    AbstractPaperGuard 测到提示注入与白盒视觉抬分，并比较分块检索防御。

    PaperGuard同时评测并检测多模态 AI 审稿中的评分操纵。

    Paper analysisRead analysis
  12. AttacksarXiv 2605.28074

    SilentRetrieval: Hijacking retrieval-augmented generation with semantically preserved adversarial data poisoning

    提出 SilentRetrieval，以语义保持对抗文档投毒劫持 RAG 检索与生成

    arXiv
    2605.28074
    Published
    Layer
    Application layer
    Setting
    LLM applications
    Evaluated models
    Llama-2-7B-Chat, Mistral-7B-Instruct-v0.2, Qwen-7B-Chat and others10
    Abstract两阶段投毒在单文档协议下兼顾检索劫持、生成操纵与近良性困惑度。

    SilentRetrieval 是针对 RAG 的两阶段语料投毒，在每查询一篇、合成目标答案的统一协议下，同时处理检索劫持、生成操纵和流畅度。

    Paper analysisRead analysis
  13. AttacksarXiv 2602.00364

    "Someone hid it": Query-agnostic black-box attacks against LLM retrieval

    提出查询无关黑盒攻击 DQ-A，用可迁移文末后缀降低文档检索排名

    arXiv
    2602.00364
    Published
    Layer
    Application layer
    Setting
    LLM applications
    Evaluated models
    Qwen-1.5-7B-Instruct, SFR-Embedding-Mistral, E5-Mistral-7B-Instruct and others10
    AbstractDQ-A 以无查询黑盒后缀压低 LLM 检索排名，并在 BRIGHT 上比较三种基线。

    DQ-A 是一种 query-agnostic、complete black-box 的文档隐藏攻击，对象是用最后隐层相似度匹配的 LLM 检索。

    Paper analysisRead analysis
  14. AttacksarXiv 2512.01335

    EmoRAG 研究揭示 RAG 检索对表情符号扰动的脆弱性

    提出 EmoRAG，向查询与知识库注入同一表情以劫持 RAG 检索

    arXiv
    2512.01335
    Published
    Layer
    Application layer
    Setting
    LLM applications
    Evaluated models
    SPECTER, Contriever, Qwen2-7B and others11
    AbstractEmoRAG 用表情联动查询与库文本。

    EmoRAG 测量 RAG 对符号扰动的反应：同一表情同时出现在查询和库文本里时，检索会偏离语义匹配。作者要检查的是，检索是否真由语义相关性决定。

    Paper analysisRead analysis
You have reached the end