Skip to content

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

Oct 12 · Monday

All papers

Found 356 papers
Filter3

Attack type

Jailbreaks252Prompt injection244Poisoning and backdoors154Model tampering35Extraction and theft81Detection evasion75Denial of service10Malicious use27

Attack technique

Gradient and representation optimization18Automated search54Multi-turn escalation28Encoding and obfuscation39Role-play and persuasion24Conversation structure manipulation12Reasoning-chain manipulation7Cross-modal payloads33Side channels4Defense-targeted11

Model risks

Deception and scheming4Misalignment and value divergence3Inappropriate refusal9Hallucination and fabrication3

Defense type

Model hardening35Inference-time intervention22Detection and filtering28Instruction–data separation13Permissions and approval6Monitoring and auditing14Watermarking and provenance4

Impact

Harmful content221Unauthorized actions21Data and privacy leaks79Output manipulation14Oversight failure5

System type

Models and APIs280LLM applications17AI Agent59coding agent14science agent3web agent4Multi-agent systems7

Attacker access

Black-box136Gray-box10White-box36No-box (offline transfer)4

Attack surface / entry point

User input307Documents and messages3Tool responses17Tool and skill descriptions6Code repositories and dependencies5Retrieval corpus3Memory writes3Inter-agent messages6Images, audio and video59Training data11Weights and gradients17Inference services and hardware10

Layer

Prompt layer169Model layer71Application layer70Training and data layer37Infrastructure layer9

System component

MCP and skills6Memory4RAG11Guardrails and judges58Monitors17Reasoning chains19Permissions and sandboxing8Fine-tuning and open weights9Vision40Audio16Image generation27Embeddings5Multimodal82
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：JailbreaksAttack type：Extraction and theftAttack type：Malicious useClear all

Another 261papers 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
  2. DefensesarXiv 2610.13333

    LETHE: Defending Against Unauthorized Image-to-Video Generation via Value Diversity Collapse

    Proposes LETHE, which perturbs reference images to block unauthorized image-to-video

    arXiv
    2610.13333
    Published
    Layer
    Prompt layer
    Setting
    Models and APIs
    Evaluated models
    Wan2.1-I2V-14B, CogVideoX-5B-I2V, LightX2V and others6
    AbstractLETHE suppresses the diversity of reference values and attracts attention, making protected images difficult to animate faithfully.

    LETHE is a protective perturbation added to personal images before publication, used to counter unauthorized image-to-video generation.

    Paper analysisRead analysis
  3. InterpretabilityarXiv 2610.13078

    Attention sees, MLP refuses: tracing the safety pathway from attention to MLP in LLMs

    Trace attention to identify harmfulness to the MLP refusal write pathway and propose Sentinel

    arXiv
    2610.13078
    Published
    Layer
    Model layer
    Setting
    Models and APIs
    Evaluated models
    Llama-3.1-8B-Instruct, Qwen2.5-7B-Instruct, Qwen3-8B-Instruct and others18
    AbstractAttention updates are more readable for harmfulness, MLP updates write refusal, and Sentinel routes alerts using two-path signals.

    The authors reframe jailbreaking and excessive refusal of aligned models as a pathway problem on the residual stream from recognition to writing, and accordingly provide conditional reminders before decoding.

    Paper analysisRead analysis
  5. Evaluation and benchmarksarXiv 2610.11766

    研究提出意图恢复率作为 LLM 安全评测的补充指标

    提出 IRIS，用意图恢复率区分遮蔽越狱下无害输出的原因

    arXiv
    2610.11766
    Published
    Layer
    Prompt layer
    Setting
    Models and APIs
    Evaluated models
    GPT-4o, Claude Sonnet 4.5, Claude Opus 4.5 and others6
    Abstract无害输出的证据权重取决于任务是否被恢复。

    IRIS 把意图恢复当作中间信号，用来解释意图遮蔽提示下为何没有出现有害协助。

    Paper analysisRead analysis
  6. Evaluation and benchmarksarXiv 2610.11112

    EpiReal-Bench：商用图像生成模型虚假声明红队基准

    构建 EpiReal-Bench，评测图像生成器将虚假主张渲染为可信证据

    arXiv
    2610.11112
    Published
    Layer
    Prompt layer
    Setting
    Models and APIs
    Attacker
    Black-box
    Evaluated models
    GPT-Image-2, Nano Banana 2, Grok Imagen 2.0 and others4
    Abstract四款商用图像模型能把假主张画成可信证据。

    EpiReal-Bench 用来量商用图像生成器会不会把虚假现实主张画成可信视觉证据。

    Paper analysisRead analysis
  7. DefensesarXiv 2610.11005

    研究：叙事包装可跨语言绕过 LLM 拒答

    提出 GUISE 基准与 AXIS 训练，防御叙事包装越狱

    arXiv
    2610.11005
    Published
    Layer
    Training and data layer
    Setting
    Models and APIs
    Evaluated models
    Qwen3-1.7B, Qwen3-4B, GLM-4-9B
    AbstractGUISE 分开测包装与语域。

    作者测量叙事包装下的拒答失效，并训练一个在未见包装上拒绝、同时仍回答同风格良性请求的适配器。问题是：安全对齐模型拒绝直说的有害请求，却回答嵌进角色或叙事里的同一请求。文言文被用来把语域和包装拆开。

    Paper analysisRead analysis
  9. Evaluation and benchmarksarXiv 2610.10276

    PatchBench 提出激活补丁局部评测协议，揭示越狱修复的附带损伤

    提出 PatchBench-Local 协议，评测激活补丁修复越狱的局部附带损伤

    arXiv
    2610.10276
    Published
    Layer
    Model layer
    Setting
    Models and APIs
    Evaluated models
    Qwen2.5-3B-Instruct, Qwen2.5-14B-Instruct, gemma-3-4b-it and others8

    Abstract论文提出 PatchBench 与局部评估协议，揭示了激活补丁在全局能力无损下仍存局部良性退化。

    Paper analysisRead analysis
  10. AttacksarXiv 2610.09981

    研究：LLM 路由决策日志可推断用户医疗、自伤等敏感话题

    提出利用路由元数据推断敏感话题的侧信道攻击

    arXiv
    2610.09981
    Published
    Layer
    Infrastructure layer
    Setting
    LLM applications
    Evaluated models
    RouteLLM (BERT router), notdiamond-0001, Vela (vLLM Semantic Router)

    Abstract论文揭示并测量了 LLM 路由决策在无文本日志中的敏感话题泄露，验证了按类别公平性防御的有效性与边界。

    Paper analysisRead analysis
  14. DefensesarXiv 2610.09703

    研究揭示 Token-Pruning 对 VLM 安全性的影响并提出 SAP 缓解机制

    提出 Safety-Aware Pruning，缓解视觉 token 剪枝放大的多模态越狱

    arXiv
    2610.09703
    Published
    Layer
    Model layer
    Setting
    Models and APIs
    Attacker
    Black-box
    Evaluated models
    LLaVA-1.5-7B
    Abstract剪枝去掉背景 token 后注意力塌向恶意前景。

    这篇工作评估视觉 token 剪枝如何改变 VLM 在多模态越狱下的行为，并给出推理时防御 SAP。

    Paper analysisRead analysis
  15. DefensesarXiv 2610.09167

    Persist-3 用顺序检验识别未见越狱来源，提升 MTK 与 Prompt Guard 2 防御效果

    提出 Persist-3 序贯检验，从隐藏表示识别未见越狱源

    arXiv
    2610.09167
    Published
    Layer
    Model layer
    Setting
    Models and APIs
    Evaluated models
    Llama2-7B-chat, Qwen2.5-7B-Instruct, Gemma-2-9B-it and others4
    AbstractPersist-3 用有限窗口检验攻击流是否超出已知越狱源，检出率低于全历史 SR-DKW，但决策只依赖最近观测。

    Persist-3 是一个有限记忆的序贯检测器，用来判断攻击表示流是否已经不能由任何一个已知越狱源解释。。

    Paper analysisRead analysis
  16. Evaluation and benchmarksarXiv 2610.09033

    研究发布 ASRD 数据集，评测五款开源模型在非规范输入下的安全表现

    arXiv
    2610.09033
    Published
    Layer
    Prompt layer
    Setting
    Models and APIs
    Evaluated models
    Gemma 2 9B, Gemma 3 4B, Llama 3.1 8B and others6
    Abstract五模型、七族表面形式的四态评测里，低有害遵从不等于读懂后的安全处理。

    ASRD 用四态标签检查五个开源模型如何处理非规范表面上的有害请求。

    Paper analysisRead analysis
  17. DefensesarXiv 2610.08678

    研究揭示投机解码的安全代价并提出 SecureSD 加固方法

    测量有损投机解码的安全代价，并提出 SecureSD 收紧早期验证

    arXiv
    2610.08678
    Published
    Layer
    Model layer
    Setting
    AI Agent
    Attacker
    Black-box
    Evaluated models
    Qwen3-8B, Qwen3-32B, Llama-3.1-8B-Instruct and others6
    Abstract有损推测解码的早期偏离更快抬高越狱与注入风险。

    作者研究有损推测解码在换取速度时，是否会比效用更快地损害越狱与提示注入防护，并给出按位置校正的 SecureSD。

    Paper analysisRead analysis
  18. DefensesarXiv 2610.08316

    MARCO：面向蛋白质生成模型的放射性水印框架

    提出 MARCO 放射性水印，冻结蛋白质生成模型并在去噪时嵌入签名

    arXiv
    2610.08316
    Published
    Layer
    Model layer
    Setting
    Models and APIs
    Attacker
    Black-box
    Evaluated models
    RFdiffusion, ESMFold, Chroma and others6
    AbstractMARCO 在冻结的 PGM 上去噪时嵌构象水印，作者称能兼顾保真与抗扰动，并使盗版模型继承水印。

    MARCO 是加在蛋白质生成模型推理过程上的构象水印，用来标记知识产权，并为可能的生物安全滥用保留可追溯信号。作者还把它说成数字权利管理工具，以及符合生物武器公约的数字监管链。问题在于：专有 PGM 的三维输出可能被拿去训练盗版模型，也可能在来源被抹掉后绕过筛查。

    Paper analysisRead analysis
  19. DefensesarXiv 2610.07532

    LADE：用暗知识中的模型无关潜在安全信号防御 LLM 越狱

    提出 LADE，用首 token 暗知识在生成前过滤越狱查询

    arXiv
    2610.07532
    Published
    Layer
    Prompt layer
    Setting
    Models and APIs
    Attacker
    White-box
    Evaluated models
    Llama-2-7B-Chat, Llama-3-8B-Instruct, Qwen2-7B-Instruct and others13
    Abstract作者用 LADE 在生成前过滤有害查询，并称八个模型中六个的 Held-out 平均最好。

    LADE 是一种只读首 token 输出概率的防御，用来在生成前挡下可能引出有害回复的查询。

    Paper analysisRead analysis
  20. DefensesarXiv 2610.07276

    SAFESHIELD：小语言模型部署时安全的决策组织框架

    提出 SAFESHIELD，组织小语言模型部署时的安全决策

    arXiv
    2610.07276
    Published
    Layer
    Application layer
    Setting
    LLM applications
    Evaluated models
    Qwen3-4B-Instruct
    AbstractSAFESHIELD 用可审计决策轨迹组织 SLM 部署时的四项安全决策，并报告协调依赖影响端到端结果。

    SAFESHIELD 是面向小语言模型的部署时安全框架，把运行时安全从单项审核改写成可检查的决策过程。

    Paper analysisRead analysis
Load more