Skip to content

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

Oct 12 · Monday

All papers

Found 15 papers
Filter5

Attack type

Jailbreaks8Prompt injection4

Attack technique

Gradient and representation optimization6Automated search6Encoding and obfuscation5Role-play and persuasion4Conversation structure manipulation4Reasoning-chain manipulation14Cross-modal payloads10Defense-targeted4Dormant triggers4

Impact

Harmful content7Output manipulation5

System type

Models and APIs11

Attacker access

Black-box6

Attack surface / entry point

User input8Images, audio and video6

Layer

Prompt layer10Application layer4

System component

MCP and skills7Memory4RAG7Guardrails and judges21Reasoning chains9Fine-tuning and open weights4Vision41Audio6Image generation7Embeddings3Multimodal54
Select any value within a category; different categories must all match. Counts include the other filters you have selected.Another 247papers have no tags yet and are not included in filtered results.
Attack technique：Gradient and representation optimizationAttack technique：Encoding and obfuscationAttack technique：Conversation structure manipulationSystem component：Reasoning chainsSystem component：AudioClear all

Another 247papers 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
  3. AttacksarXiv 2609.27542

    Researchers use control token injection to delete gpt-oss-20b's reasoning chain and bypass chain-of-thought monitoring

    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
  4. 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
  5. 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
  6. Evaluation and benchmarksarXiv 2606.06037

    SpeechJBB：语码转换语音下大型音频语言模型的越狱基准

    提出 SpeechJBB，评测语码转换语音下音频语言模型的越狱稳健性

    arXiv
    2606.06037
    Published
    Layer
    Prompt layer
    Setting
    Models and APIs
    Evaluated models
    Qwen2.5-Omni-7B, Qwen3-Omni-30B-A3B-Instruct, Voxtral-Small-24B and others9
    AbstractSpeechJBB 用五语语码转换语音和伪词探针评测九个 LALM，作者将越狱归为对齐失败而非听不懂。

    SpeechJBB 用合成语音测量 LALM 在五语、语码转换和局部伪词混淆下的安全对齐，并另设口语理解对照。

    Paper analysisRead analysis
  8. AttacksarXiv 2604.14604

    The AudioHijack framework achieves covert audio prompt injection against 13 speech large models

    提出 AudioHijack，用不可感知对抗音频对语音模型做上下文无关提示注入

    arXiv
    2604.14604
    Published
    Layer
    Prompt layer
    Setting
    Models and APIs
    Evaluated models
    SpeechGPT, GLM-4-Voice, VITA-Audio and others15

    Abstract提出了 AudioHijack 框架，揭示了 LALM 对听觉提示注入的脆弱性与传统防御的局限。

    Paper analysisRead analysis
  10. AttacksarXiv 2603.10091

    多流扰动攻击：通过并发任务干扰突破思考型 LLM 的安全对齐

    提出多流扰动攻击，交织有害与良性任务绕过思考型 LLM 安全对齐

    arXiv
    2603.10091
    Published
    Layer
    Prompt layer
    Setting
    Models and APIs
    Evaluated models
    Qwen3 1.7B, Qwen3 4B, Qwen3 8B and others11
    Abstract多流扰动用任务交织测试 thinking 模式，作者称它同时冲击内容安全与推理稳定性。

    多流扰动攻击把有害任务与良性任务织进同一提示词，测试 thinking 模式的内容安全与推理稳定性。作者认为，交织多任务会拉长思考，使安全检测面对被打散的有害意图，并可能把推理用于合理化，或陷入崩溃与重复。方法不访问梯度。

    Paper analysisRead analysis
  11. AttacksarXiv 2602.18514

    招聘中的特洛伊木马：标准模型与推理模型间接提示注入的红队测试案例研究

    用简历中的不可见指令做间接提示注入，操纵招聘推荐

    arXiv
    2602.18514
    Published
    Layer
    Application layer
    Setting
    LLM applications
    Evaluated models
    Qwen 3 30B A3B Instruct 2507, Qwen 3 30B A3B Thinking 2507

    Abstract作者称推理在简单注入下会帮攻击圆谎，在不合逻辑的复杂注入下又会把欺骗漏出来。

    Paper analysisRead analysis
  14. AttacksarXiv 2504.01094

    研究者提出 Multi-AudioJail

    提出 MULTI-AUDIOJAIL，以多语多口音和声学扰动越狱音频模型

    arXiv
    2504.01094
    Published
    Layer
    Prompt layer
    Setting
    Models and APIs
    Attacker
    Black-box
    Evaluated models
    Qwen2-Audio, DiVA-llama-3-v0-8b, MERaLiON-AudioLLM-Whisper-SEA-LION and others5

    Abstract黑盒多语多口音音频加声学扰动可抬高 LALM 的 JSR，文本示范防御并非在每个模型上都下降。

    Paper analysisRead analysis
  15. AttacksarXiv 2406.12814

    论文提出 ARE 框架，在 VisualWebArena 上评估多模态 Agent 的对抗鲁棒性

    提出 ARE 框架与图像文本攻击，评测多模态网页 Agent 的对抗鲁棒性

    arXiv
    2406.12814
    Published
    Layer
    Application layer
    Setting
    web agent
    Evaluated models
    GPT-4o, GPT-4V, Claude-3-Opus and others4

    Abstract系统评估多模态网页智能体的脆弱性，指出推理时计算组件被攻破会加剧系统风险。

    Paper analysisRead analysis
You have reached the end