Translation in progress Some entries still include text in the original language.
Oct 12 · Monday
All papers Found 26 papers Another 265 papers do not have category tags yet and are not included in filtered results.
“
About this collection Collection in progress We are building a large AI safety paper collection to help researchers find work related to their interests. Papers are collected automatically andfiltered only for relevance to AI safety , without judging quality, novelty or reading value. The default “Annotated first” shows papers with tags or published paper analyses first. Each group is ordered by original publication date, newest first, with undated papers last. “Newest first” uses publication dates only. This order is not a quality ranking.
One-line summaries, tags and paper analyses are prepared by AI from the papers. Language models may have systematic biases and produce hallucinations.We plan to carry out human annotation and review to reduce errors in this content.
提出 TRANSCOPE，用 CPU 硬件计数器推断训练集成员
Published Oct 6
Evaluated models BERT (4.4M-340M) , ViT (22M-307M) , Pythia (70M-2.8B) and others5 Abstract TRANSCOPE 揭示了分词器在嵌入表引发的硬件访问局部性差异，证实微架构计数器可突破恒定时间与置信度遮蔽防御。
提出 PrivDrift 基准，审计话题漂移后的用户秘密复述
Published Sep 25
Evaluated models GPT-OSS-120B , DeepSeek-R1 , Qwen3-VL-235B 消融比较方言越狱中表层形式、文化框架与策略库的作用
Published Sep 15
Evaluated models GPT-4o , DeepSeek-Reasoner , DeepSeek-Chat 提出 SAST-IR，用有状态攻击者迭代说服无记忆模型接受反事实
Published Sep 15
Evaluated models DeepSeek-Chat 提出内容不变样式包装，翻转安全评审器的判定
Published Sep 8
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。
提出 MIPC 与 CFA，观察模型自拟测验后接受开发者身份
Published Sep 3
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。
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
Published Sep 2
Evaluated models Qwen3-Next-80B , DeepSeek-V3.2 , GLM-4.7 and others6 Abstract BLUEPRINT 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。
提出 KeyPooling，测量 API 中继何处瓦解提示缓存隔离
Published Aug 18
Evaluated models grok-4-fast-non-reasoning , grok-4-1-fast-non-reasoning , grok-3 and others11 Abstract 共享上游凭证使中继客户进入同一缓存域，保留身份或独立 principal 可分开。KeyPooling 测量 LLM API 中继在何处丢掉客户身份，使 prompt cache 不再按下游客户隔离。
提出 RolePlay 框架，用人设条件放大 LLM 单次推理开销
Published Jul 29
Evaluated models DeepSeek-V4-Pro , Gemini-3.5-Flash , Qwen-3.5-Plus and others5 提出 Concept2Scenario，借助内部概念归因发现可迁移的越狱场景
Published Jul 26
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。
评测多语言与编码混淆提示下模型生成钓鱼与键盘记录内容的遵从
Published Jun 29
Evaluated models DeepSeek-V3.2 , Gemini 2.5 Flash , GPT-5 Mini and others6 提出 Ghostwriter 攻击，用伪权威证据条件注入操纵模型观点
Published Jun 4
Evaluated models GPT-4o , Claude-3.7-Sonnet , DeepSeek-V3 and others10 Abstract 作者称攻击未被消除。Ghostwriter 是一种针对上下文窗口的观点引导攻击：把误导陈述改写成伪权威文本，再在相关查询中让目标模型把该观点写进回答。
构建 MultiTurnPSB，评测四轮医疗越狱与分类器输入防御
Published May 30
Evaluated models GPT-4.1-mini , Claude Sonnet 4.5 Abstract 多轮攻击加剧医疗模型风险，分类器干预有效但受高误报率制约。MultiTurnPSB 针对患者端医疗大模型在持续交互中的安全风险，构建了包含三种攻击模态的四轮对抗评测基准。
提出 Persona Attack，分步写入对话记忆指令以越狱
Published May 29
Evaluated models GPT-4o , Llama-3.2-3B-Instruct , ChatGPT-4o and others5 提出 PCSA，用人设多轮模拟咨询来访者以诱导有害共情回复
Published Apr 7
Evaluated models Llama-3.1-8B , Llama-3.1-70B , GPT-3.5-Turbo and others8 提出复合越狱，叠对比结构、权威人设与自评饱和指令层级
Published Apr 3
Evaluated models OpenAI gpt-oss-20b 提出 FoodGuardBench 与护栏 FoodGuard-4B，评测并降低食品危害越狱
Published Apr 2
Evaluated models Claude-3.7-Sonnet , GPT-4o , GPT-4.1 and others10 提出 S2C 越狱，用情境改写、切分和线索伪装推迟有害意图整合
Published Mar 17
Evaluated models Llama3.3-70B-Instruct , DeepSeek-r1 , DeepSeek-v3 and others16 提出 STAR 多轮角色扮演框架，利用对话状态演化绕过安全对齐
Published Mar 15
Evaluated models GPT-4o , Claude 3.5 Sonnet , Gemini 2.0-Flash and others5 提出 OPTILEAK，经 KV 缓存侧信道逐 token 重建他人提示
Published Feb 24
Evaluated models Qwen-2.5-3B-Instruct , Qwen-2.5-7B-Instruct , Qwen-2.5-14B-Instruct and others4 Abstract OPTILEAK 用 SFT 与 hard-token DPO 降低每 token 请求数，作者称缓存提示泄露的风险高于先前报告。