Skip to content

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

Oct 12 · Monday

All papers

Found 86 papers
Filter4

Attack type

Jailbreaks9Prompt injection68Poisoning and backdoors83Extraction and theft4Detection evasion45Denial of service9Malicious use3

Attack technique

Gradient and representation optimization5Automated search10Defense-targeted7Dormant triggers24Selection manipulation5

Defense type

Model hardening7Inference-time intervention3Detection and filtering13Instruction–data separation3

Impact

Harmful content15Unauthorized actions48Data and privacy leaks19Output manipulation75Availability and reliability11Oversight failure6

System type

Models and APIs40LLM applications18AI Agent28coding agent4Multi-agent systems4

Attack surface / entry point

User input16Web pages and interfaces6Documents and messages7Tool responses5Tool and skill descriptions5Code repositories and dependencies3Retrieval corpus21Memory writes8Images, audio and video7Training data37Weights and gradients15Distribution channels14

Layer

Prompt layer4Model layer8Application layer42Training and data layer31

System component

MCP and skills5Memory14RAG20Guardrails and judges4Reasoning chains4Fine-tuning and open weights16Vision8Embeddings3Multimodal13
Select any value within a category; different categories must all match. Counts include the other filters you have selected.Another 261papers have no tags yet and are not included in filtered results.
Attack type：Poisoning and backdoorsAttack type：Malicious useImpact：Output manipulationImpact：Availability and reliabilityClear all

Another 261papers do not have category tags yet and are not included in filtered results.

  1. Analysis and theoryarXiv 2610.13051

    Space beats statistics, retraining beats ranking: auditing training labels under label-only backdoors

    Auditing training labels for label-only backdoors, comparing spaces and statistics, and judging by retraining

    arXiv
    2610.13051
    Published
    Layer
    Training and data layer
    Setting
    Models and APIs
    Evaluated models
    ResNet-32
    AbstractLabel inspection must be done in a space that has not been fitted to these labels, and judged by retrained models rather than detection scores.

    The authors ask: can a defender without trusted data decide which training labels to keep, and how much clean data this destroys, and what the final model accuracy is.

    Paper analysisRead analysis
  2. AttacksarXiv 2610.12644

    Cancerous backdoor attack: progressive activation via invisible tokens and a stage controller.

    Proposes the Cancerous backdoor, which progressively corrupts reasoning using invisible stage tokens.

    arXiv
    2610.12644
    Published
    Layer
    Training and data layer
    Setting
    Models and APIs
    Evaluated models
    Llama-3.1-8B-Instruct, Qwen2.5-7B-Instruct, Mistral-7B-Instruct-v0.3
    AbstractHidden tokens plus a stage controller make the backdoor progressively corrupt reasoning upon triggering, and the authors believe that looking only at the answer will miss it.

    Cancerous Backdoor Attack changes LLM backdoors from one-time activation to multi-stage reasoning corruption that accumulates progressively with triggers.

    Paper analysisRead analysis
  3. AttacksarXiv 2610.11932

    A Fudan team measures the citation fragility of AI search: ordinary posts can enter citations and answer text

    Measures citation fragility in AI search, demonstrating that ordinary public posts can enter citations and answers.

    arXiv
    2610.11932
    Published
    Layer
    Application layer
    Setting
    LLM applications
    Evaluated models
    Grok, Doubao, ChatGPT and others10

    AbstractThe authors claim the paper quantifies vulnerabilities in AI search caused by citation preferences and low posting thresholds; tests show public web posts can penetrate AI citations, and they call for auditing the citation supply chain.

    Paper analysisRead analysis
  5. DefensesarXiv 2610.09941

    OrthoPurify 用一步正交投影净化被后门污染的视觉语言模型

    提出 OrthoPurify，用正交投影清除视觉语言模型适配器中的劫持方向

    arXiv
    2610.09941
    Published
    Layer
    Training and data layer
    Setting
    Models and APIs
    Evaluated models
    LLaVA-1.5-7B, Qwen3-VL-8B, LLaVA-1.5-13B and others5
    Abstract一次投影去掉劫持方向。

    OrthoPurify 是部署前的 LVLM 后门净化方法，用一次正交投影去掉被劫持的权重更新方向。

    Paper analysisRead analysis
  6. AttacksarXiv 2610.09819

    FAB 攻击：给 HuBERT/WavLM 植入后门，背景警笛声即可让下游语音任务失灵

    提出 FAB 攻击，向声学基础模型植入可由环境声触发的后门

    arXiv
    2610.09819
    Published
    Layer
    Training and data layer
    Setting
    Models and APIs
    Evaluated models
    HuBERT-base, WavLM-base

    Abstract论文提出了针对声学基模的 FAB 后门攻击，在弱威胁模型下实现了任务无关、物理可触发的后门植入。

    Paper analysisRead analysis
  7. AttacksarXiv 2610.09264

    研究者提出 PackHallu：经规则文件提示注入对编码 Agent 发起包幻觉攻击

    提出 PackHallu，经规则文件提示注入诱导编码 Agent 换用攻击者指定包

    arXiv
    2610.09264
    Published
    Layer
    Application layer
    Setting
    coding agent
    Evaluated models
    Qwen2.5-Coder-7B, Qwen3-Coder-30B, Devstral-Small-2-24B and others19
    AbstractPackHallu 诱导规则文件换包。

    PackHallu 研究一种针对编程智能体规则文件的提示注入：让智能体在本应使用合法依赖时，改导入攻击者指定并控制的包。

    Paper analysisRead analysis
  8. AttacksarXiv 2610.07645

    SkillPoison：用成功经验投毒自进化 Agent 的技能形成

    提出 SkillPoison，用局部有效的成功经验诱导自进化 Agent 形成过度泛化技能

    arXiv
    2610.07645
    Published
    Layer
    Application layer
    Setting
    AI Agent
    Evaluated models
    deepseek-v4-flash, Qwen V3.8 Flash, GPT-5.5
    AbstractSkillPoison 通过局部有效且验证成功的经验诱导技能过度泛化，在多个评测基准上实现了持久的下游攻击影响。

    SkillPoison 提出了一种针对自进化智能体经验到技能形成过程的攻击范式，通过组织已验证的成功经验诱导技能过度泛化。

    Paper analysisRead analysis
  9. AttacksarXiv 2610.06049

    研究者提出仅修改 SAE 解码器的后门攻击，可在不改动语言模型的情况下植入行为

    提出只改 SAE 解码器的后门，在冻结语言模型上植入代码插入

    arXiv
    2610.06049
    Published
    Layer
    Model layer
    Setting
    Models and APIs
    Evaluated models
    Gemma 3 4B Instruct, Qwen3 1.7B, Llama 3.1 8B Instruct
    Abstract只改解码器的 SAE 可带上代码插入后门，且若干常规质量指标变化不大。

    只改 SAE 解码器、冻结编码器与语言模型，即可把代码插入行为放进推理时的辅助组件。作者把这视为 SAE 供应链问题：checkpoint 一旦替换某一层激活，就参与后续计算，而不只是解释工具。

    Paper analysisRead analysis
  10. DefensesarXiv 2610.05826

    研究者提出临时可见性协议，限制持续摄入 RAG 向量库的投毒暴露

    提出临时可见性协议，限制持续摄入 RAG 向量库的投毒暴露

    arXiv
    2610.05826
    Published
    Layer
    Application layer
    Setting
    LLM applications
    Abstract用截止期限制未核验可见时长。

    这篇工作给持续摄入的 RAG 向量库一个失败关闭的临时可见性协议，把协议能限制的未核验暴露和依赖检测器的中毒暴露分开。

    Paper analysisRead analysis
  11. AttacksarXiv 2610.05793

    研究者训练出按多智能体拓扑触发的代码后门模型

    提出拓扑条件后门，使代码模型在推断多智能体部署时植入漏洞

    arXiv
    2610.05793
    Published
    Layer
    Training and data layer
    Setting
    coding agent
    Evaluated models
    Qwen2.5-7B-Instruct, Qwen2.5-Coder-7B-Instruct

    Abstract论文证实了拓扑条件后门的可训练性与泛化性，展示了常规评测盲区，并提出差异化拓扑审计作为检测防御方向。

    Paper analysisRead analysis
  12. Analysis and theoryarXiv 2610.02886

    研究揭示无触发投毒审计缺口：事实答对不等于决策正确

    揭示无触发虚假训练下事实回答与下游决策的审计差距

    arXiv
    2610.02886
    Published
    Layer
    Training and data layer
    Setting
    Models and APIs
    Evaluated models
    Qwen2.5-0.5B, Qwen2.5-1.5B, Qwen2.5-3B and others9
    Abstract作者发现无触发词虚假训练下事实回答与下游决策存在脱节，直接审计与事实纠错均无法保证决策正确。

    论文是一项针对语言模型在无触发词虚假信息训练下行为表现的实证分析研究。

    Paper analysisRead analysis
  16. DefensesarXiv 2609.39548

    NDDL：用正常扩散动力学为文生图模型做后门防御

    提出 NDDL，用良性扩散轨迹检测文生图后门提示并定位触发 token

    arXiv
    2609.39548
    Published
    Layer
    Model layer
    Setting
    Models and APIs
    Evaluated models
    Stable Diffusion v1.5, Stable Diffusion XL, Stable Diffusion v3.5 and others4
    AbstractNDDL 从良性扩散轨迹学习正常转移，用动态不一致检测后门并定位触发 token。

    NDDL 是面向文生图扩散模型的后门防御：防御者只有白盒模型访问和良性提示词，没有后门先验，需要区分后门提示词并找出触发 token。

    Paper analysisRead analysis
  17. Evaluation and benchmarksarXiv 2609.39146

    MADBench：多智能体辩论安全性评测基准发布

    提出 MADBench，评测多智能体辩论是吸收还是放大对抗影响

    arXiv
    2609.39146
    Published
    Layer
    Application layer
    Setting
    Multi-agent systems
    Evaluated models
    gpt-4o, GPT-5, GPT-6 Astra
    AbstractMADBench 显示辩论能吸收部分准确率攻击，同时放大未授权读写。

    MADBench 是一个评测多智能体辩论安全性的基准，用来看攻击是被辩论吸收还是被放大。

    Paper analysisRead analysis
  19. DefensesarXiv 2610.00348

    NEEDLE：通过权重正交化移除 LLM 后门

    提出 NEEDLE，用权重正交化在已知触发器时去除大模型后门

    arXiv
    2610.00348
    Published
    Layer
    Model layer
    Setting
    Models and APIs
    Evaluated models
    Gemma-3-4B-IT, Qwen3-4B-Instruct-2507, Gemma-3-12B-IT

    AbstractNEEDLE 用闭式权重编辑去掉已知触发器的后门，并保住拒绝子空间。

    Paper analysisRead analysis
  20. AttacksarXiv 2609.38270

    VirusCascade：劫持 LLM 推荐智能体的协同反思机制

    提出 VirusCascade，联合操纵商品描述与交互轨迹以劫持推荐智能体协同反思

    arXiv
    2609.38270
    Published
    Layer
    Application layer
    Setting
    Multi-agent systems
    Evaluated models
    AgentCF, AgentSEQ, AgentRAG and others4
    Abstract作者把反思写回视为推广攻击入口，并用联合操纵和诊断探针研究其持续与传播。

    VirusCascade 是针对推荐智能体协作反思的 black-box targeted promotion 攻击研究。 攻击者只改商品描述和少量账号的普通交互，试图让目标商品获得更多良性用户曝光。

    Paper analysisRead analysis
Load more