Another 227papers do not have category tags yet and are not included in filtered results.
提出 CodeMimicry，以结构化代码补全诱导模型生成有害内容
提出 CollageAttack，用空间文本碎片重组越狱文生图模型
Abstract系统揭示 T2I 跨模态空间合成安全漏洞，证实离散无害元素可在图像层面重构有害意图并放大传播强度。
提出任意字母置换密码越狱，无需微调即可诱导有害输出
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。
提出 DIVA，将有害意图拆成视觉锚点与运动文本以越狱视频生成模型
DIVA 是针对带参考图的视频生成模型的免训练 black-box 越狱：用一致性机制本身堵住语义漂移这条安全逃逸路径。。
提出 RoguePrompt 双层编码自重构越狱，绕过 LLM 审核
提出 OTTER 词元替换越狱，绕过表面毒性审核诱导有害回答
Abstract论文展示了表面毒性与恶意意图的解耦现象，通过标准 API 实现黑盒越狱测试，并提出闭环分类器加固建议。
提出 JailbreakOPT，组合原子工具迭代优化黑盒单轮越狱提示
提出 HPAA 攻击，用排版视觉线索绕过内容审核系统
这篇论文研究了人类视觉感知与大模型分词处理机制之间的差异所导致的内容审核漏洞。
提出 DSR 跨模态越狱，用无害图文诱导 MLLM 生成有害图像
DSR 是一种 black-box 跨模态越狱：输入侧不放显式有害内容，而让商业 MLLM 在推理中把无害图文重组成有害图像。
提出 Babel 混淆分布采样，黑盒越狱闭源 LLM 安全注意力
提出多图越狱框架 DMN，把有害意图拆进图像序列绕过安全护栏
提出删字变体越狱方法，利用重构与隐藏权衡绕过多模态模型安全过滤
提出 Formal Logic 等数学编码越狱，绕过自然语言安全过滤
作者在 black-box、只提交纯文本的设定下，比较六种数学编码能否让仍开着安全措施的模型产出有害内容。
Extract system instructions with encoding and structured requests, and harden them by rewriting with CoT
This work evaluates and hardens the confidentiality of LLM system instructions under encoded extraction. The authors are concerned with: when direct extraction has been rejected, whether the model still surrenders protected content when the same request is rewritten as a formatting or encoding task.
Propose S2C jailbreak, which delays harmful intent integration using contextual rewriting, fragmentation, and clue-guided camouflage.
Propose StructAttack, which uses structural diagrams to induce visual models to assemble malicious content from harmless semantic slots.
StructAttack is a black-box, single-query jailbreak against LVLMs: it decomposes a harmful instruction into separate seemingly harmless slots, draws them into a structural diagram, and then makes the model reassemble the slots into harmful content.
Propose TFM, using start and end frames and implicit substitution to jailbreak text-to-video models
TFM is a strictly black-box prompt jailbreak for text-to-video that exploits the model's temporal infilling of unspecified intermediate processes.
提出 MIDAS ，把有害语义拆进多张视觉谜题并跨图重构以越狱多模态模型
Propose CC-BOS, which uses fruit fly search in an eight-dimensional policy space to optimize classical Chinese jailbreak prompts
Propose COMET cross-modal entanglement attack to jailbreak vision-language models
COMET is a black-box jailbreak for VLMs with multimodal reasoning, used to red-team cross-modal safety alignment. The gap the authors address is: distributing malicious cues into fixed combinations of image and text cannot detect vulnerabilities that only appear in multi-step reasoning. The attack does not access gradients; it only constructs image and text inputs.