Skip to content

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

Oct 13 · Tuesday

All papers

Found 4 papers
Filter4

Attack technique

Gradient and representation optimization4Automated search6Multi-turn escalation5Encoding and obfuscation13Conversation structure manipulation0Cross-modal payloads24

System type

Models and APIs4

Attack surface / entry point

User input4Web pages and interfaces4Images, audio and video25

Layer

Prompt layer3

System component

Guardrails and judges4Reasoning chains4Vision4Multimodal4
Select any value within a category; different categories must all match. Counts include the other filters you have selected.Another 215papers have no tags yet and are not included in filtered results.
Attack technique：Gradient and representation optimizationAttack technique：Conversation structure manipulationAttack surface / entry point：User inputSystem component：VisionClear all

Another 215papers do not have category tags yet and are not included in filtered results.

  1. AttacksarXiv 2610.09027

    P-VMI：对抗图像经 KV cache 持续影响多轮对话

    提出持久视觉记忆注入，使对抗图像经缓存在触发时操控输出

    arXiv
    2610.09027
    Published
    Layer
    Infrastructure layer
    Setting
    Models and APIs
    Evaluated models
    Qwen3-8B, Qwen3-VL-8B-Instruct, Gemma-4-12B

    Abstract论文提出 P-VMI，证实对抗输入被掩蔽后仍可经 KV 缓存持久控制模型，强调了保留状态的安全风险。

    Paper analysisRead analysis
  3. AttacksarXiv 2601.06464

    Adversarial Robustness of 3D Large Vision-Language Models

    提出 Vision 与 Caption 攻击，扰动点云操纵 3D VLM 输出

    arXiv
    2601.06464
    Published
    Layer
    Prompt layer
    Setting
    Models and APIs
    Evaluated models
    PointLLM-7B, GPT4Point-2.7B
    Abstract作者用两类白盒攻击评估 3D VLM，称无目标更易成功、定向更难。

    对 point-based 3D VLM 的白盒对抗鲁棒性评测。 作者关心驾驶、机器人等场景里，小几何扰动能否改写类别或描述，以及定向诱导特定输出是否像 2D 那样容易。方法上，Vision Attack 用余弦相似度推远原视觉 token，或拉向另一点云的特征。

    Paper analysisRead analysis
  4. AttacksarXiv 2511.16110

    MFA 框架在 17 个视觉语言模型上取得 58.5% 攻击成功率

    提出 MFA 框架，协同突破视觉语言模型的对齐、提示与审核层

    arXiv
    2511.16110
    Published
    Layer
    Prompt layer
    Setting
    Models and APIs
    Attacker
    Black-box
    Evaluated models
    MiniGPT-4, LLaMA-4-Scout-Inst, LLaMA-3.2-11B-Vision-Instruct and others17

    Abstract作者称多层防御的 VLM 仍存在跨模型漏洞，MFA 整合三项技术实现了端到端越狱。

    Paper analysisRead analysis
You have reached the end