Skip to content

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

Oct 12 · Monday

All papers

Found 54 papers
Filter3

Attack type

Jailbreaks32Prompt injection10Extraction and theft4Detection evasion5Denial of service4

Attack technique

Gradient and representation optimization63Automated search113Multi-turn escalation30Encoding and obfuscation62Role-play and persuasion32Conversation structure manipulation16Reasoning-chain manipulation13Cross-modal payloads57Side channels4Defense-targeted44Dormant triggers55Selection manipulation12

Impact

Harmful content29Data and privacy leaks3Output manipulation15Availability and reliability4

System type

Models and APIs39LLM applications8AI Agent7coding agent3

Attacker access

Black-box28White-box3

Attack surface / entry point

User input43Documents and messages6Tool responses3Code repositories and dependencies4Retrieval corpus3

Layer

Prompt layer36Application layer14

System component

Guardrails and judges10Monitors3Reasoning chains15Multimodal3
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.
Attack technique：Role-play and persuasionAttack technique：Conversation structure manipulationAttack technique：Reasoning-chain manipulationClear all

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

  1. 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
  2. AttacksarXiv 2610.04985

    研究揭示 Jev 作为应用决策层的安全与隐私风险

    评估应用决策层的输入注入与隐私推断风险

    arXiv
    2610.04985
    Published
    Layer
    Application layer
    Setting
    LLM applications
    Attacker
    Black-box, Gray-box
    Evaluated models
    Jev, NanoJev
    AbstractJev 的类型化决策可被输入与训练操纵，也会泄露信息，并同时支持检测与滥用。

    Jev 是 TypeSafe 的 System One 模型，把应用 state 和固定问题变成 Choice、Noul 或 Score 的类型化决策与概率。作者用官方 API 和可改训练的 NanoJev，考察它作为决策层时的操纵、泄露与双重用途。

    Paper analysisRead analysis
  3. AttacksarXiv 2609.36956

    研究者提出 Controlled Decoding Attacks，通过纯文本接口越狱黑盒 LLM

    提出 BLINDBIAS，通过纯文本续写与前缀控制越狱黑盒模型

    arXiv
    2609.36956
    Published
    Layer
    Prompt layer
    Setting
    Models and APIs
    Attacker
    Black-box
    Evaluated models
    GLM-5, Gemini-3.5-Flash, Qwen3-32B and others4
    Abstract作者认为，隐藏数值概率并不足以消除支持重复采样和前缀续写的解码攻击面。

    BLINDBIAS 是一种针对 sample-only 续写接口的 black-box 解码时越狱框架，研究没有目标数值概率时，采样能否支持逐步控制。

    Paper analysisRead analysis
  6. AttacksarXiv 2609.27542

    研究者用控制 token 注入删除 gpt-oss-20b 推理链并绕过思维链监控

    提出控制标记注入，清空工具 Agent 的思维链以绕过推理监控

    arXiv
    2609.27542
    Published
    Layer
    Application layer
    Setting
    AI Agent
    Evaluated models
    openai/gpt-oss-20b, google/gemma-4-26B-A4B-it, deepseek-ai/DeepSeek-R1-Distill-Qwen-7B and others4

    Abstract论文阐明控制标记注入可抑制 CoT 使监控失效并促成拒绝绕过，且工具执行受解析器宽容度影响。

    Paper analysisRead analysis
  7. AttacksarXiv 2609.19722

    ALIBI：向二进制注入虚假安全叙事可让 LLM 恶意软件分析器误判

    提出 ALIBI，向恶意二进制注入不执行掩护叙事使分析器误判为良性

    arXiv
    2609.19722
    Published
    Layer
    Application layer
    Setting
    LLM applications
    Attacker
    Black-box
    Evaluated models
    Gemini 2.5 Pro, GPT-5.5 Pro, Claude Opus 4.7
    AbstractALIBI 用二进制内的掩护叙事重释可疑静态证据，Gemini 上翻转多，另两模型主要为降级。

    ALIBI 是针对 LLM 恶意软件分诊的语义掩护叙事攻击：不覆盖系统指令，而把仍可见的可疑静态证据重释成良性安全产品的预期行为。

    Paper analysisRead analysis
  10. AttacksarXiv 2609.15989

    研究者提出 plan injection 攻击，可绕过思维链监控

    提出 plan injection 攻击，植入良性伪装计划以绕过思维链监视器

    arXiv
    2609.15989
    Published
    Layer
    Prompt layer
    Setting
    Models and APIs
    Attacker
    Gray-box
    Evaluated models
    Qwen2.5-Coder-7B-Instruct, Qwen3-8B, DeepSeek-R1-0528 and others6

    Abstract论文指出模型会将外部上下文中的伪装计划改写为自身推理，使思维链监视器失效甚至反向合理化恶意行为。

    Paper analysisRead analysis
  11. AttacksarXiv 2609.09553

    研究：任意密文攻击前沿大模型无需微调即可越狱

    提出任意字母置换密码越狱，无需微调即可诱导有害输出

    arXiv
    2609.09553
    Published
    Layer
    Prompt layer
    Setting
    Models and APIs
    Attacker
    Black-box
    Evaluated models
    Claude Sonnet 4, Claude Sonnet 4.5, Gemini 3 Flash (preview) and others7

    Abstract前沿大模型无需微调即可通过提示词学会字母置换密码，进而绕过分类器与对齐，作者呼吁关注能力提升带来的安全风险。

    Paper analysisRead analysis
  13. AttacksarXiv 2609.08236

    研究：内容不变的风格包装可翻转 LLM 安全评判器的判定

    提出内容不变样式包装，翻转安全评审器的判定

    arXiv
    2609.08236
    Published
    Layer
    Prompt layer
    Setting
    Models and APIs
    Attacker
    Black-box
    Evaluated models
    GPT-4o-mini (gpt4o), GPT-4o-mini (strongreject), DeepSeek-Chat and others7
    Abstract论文揭示特定安全评审器易受保持内容不变的外层包装诱导改变判定，并指出安全评估应将判定稳定性作为重要考量指标。

    这是一项关于大语言模型自动化安全评审系统表面样式脆弱性的评估研究。

    Paper analysisRead analysis
  14. Risky behaviorarXiv 2609.08186

    论文揭示大推理模型的对齐崩溃并提出 Reasoning Trap 越狱范式

    揭示推理加深引发的对齐坍塌，用 Reasoning Trap 放大越狱

    arXiv
    2609.08186
    Published
    Layer
    Model layer
    Setting
    Models and APIs
    Evaluated models
    Qwen3-8B, Qwen3-14B, Llama2-7B and others8
    Abstract更深推理提升题目准确率，同时提高扰动下的相对损失。

    作者研究扩展推理是否以对齐稳健性为代价，提出 ALR、RT 与 RRA。。

    Paper analysisRead analysis
  15. AttacksarXiv 2609.03247

    研究揭示大语言模型的自发身份认证失败：五款模型的分阶段开发者身份实验

    提出 MIPC 与 CFA，观察模型自拟测验后接受开发者身份

    arXiv
    2609.03247
    Published
    Layer
    Prompt layer
    Setting
    Models and APIs
    Evaluated models
    ChatGPT, Claude, Qwen3.7-Plus and others5
    Abstract自拟测验可造成对话层虚假认证，但所测授权边界并未因此改变。

    这项工作把“模型自己出题、自己批改、再宣布身份已核验”定义成对话层认证失败，并与后端提权分开。

    Paper analysisRead analysis
  16. AttacksarXiv 2609.02414

    Blueprint 框架用世界观模拟与 18 个影响因子提升多轮越狱成功率

    提出多轮越狱框架 BLUEPRINT，用世界观模拟与心理因素诱导有害输出

    arXiv
    2609.02414
    Published
    Layer
    Prompt layer
    Setting
    Models and APIs
    Attacker
    Black-box
    Evaluated models
    Qwen3-Next-80B, DeepSeek-V3.2, GLM-4.7 and others6
    AbstractBLUEPRINT 解耦多轮越狱策略与情境模拟，揭示任务具象化是摆脱拒绝的核心机制，为多轮安全防御提供新视角。

    BLUEPRINT 是一个将因子化社会影响力策略空间与跨轮世界观模拟解耦的多轮大语言模型安全评测框架。

    Paper analysisRead analysis
  17. AttacksarXiv 2609.01023

    Akrasia：针对推理型代码 LLM 的隐蔽后门攻击

    提出隐蔽后门攻击 AKRASIA，在推理代码模型提示中植入触发器并伪装思维链

    arXiv
    2609.01023
    Published
    Layer
    Prompt layer
    Setting
    coding agent
    Evaluated models
    Claude-Sonnet-5, GPT-5.5, Gemini-3.5-flash and others7

    Abstract论文针对推理代码大模型提出了隐蔽推断期后门攻击 AKRASIA，揭示了模型思维链可被利用进行欺骗性伪装的安全隐患。

    Paper analysisRead analysis
  18. AttacksarXiv 2608.09867

    Researchers steal proprietary LLM inference chains via cross-model compatibility of encrypted inference blocks

    利用加密思维块的跨模型重放，诱导弱模型逐字转写专有推理链

    arXiv
    2608.09867
    Published
    Layer
    Prompt layer
    Setting
    Models and APIs
    Attacker
    Black-box
    Evaluated models
    Claude Opus 4.8, Claude Sonnet 4.6, Claude Haiku 4.5 and others13

    Abstract论文揭示了客户端加密思维块跨模型互换漏洞，提出了低成本提取方法并给出防御方案。

    Paper analysisRead analysis
  20. AttacksarXiv 2607.23496

    Concept2Scenario：用 SAE 概念归因发现可迁移的越狱场景

    提出 Concept2Scenario，借助内部概念归因发现可迁移的越狱场景

    arXiv
    2607.23496
    Published
    Layer
    Prompt layer
    Setting
    Models and APIs
    Attacker
    White-box, Black-box
    Evaluated models
    Qwen3.5-9B, Llama-3.1-8B, Ministral-3-8B and others6

    Abstract该研究从表征空间因果归因揭示场景削弱拒绝的机制，发现可复用脆弱场景先验，提升多模型黑盒越狱效能与探索速度。

    Paper analysisRead analysis
Load more