Translation in progress Some entries still include text in the original language.
Oct 12 · Monday
All papers Found 14 papers Another 232 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.
固定有害指令，单轴归因图像到文本越狱的图像属性
Published Oct 4
Evaluated models GPT-4o , Claude Sonnet 4.5 , Qwen3-VL-8B-Instruct and others6 Abstract 固定文本的图像越狱拆解，只在 Qwen 上给出一条有界相关性归因。
遮挡审计 AdvWave-P 音频越狱的频带必要性与解码层关联
Published Oct 4
Evaluated models Qwen2-Audio-7B-Instruct , LLaMA-Omni Abstract AdvWave-P 的八频带排序与能量相关。这是对加性音频越狱 AdvWave-P 的频率与解码器深度审计，受害模型主要是 Qwen2-Audio-7B-Instruct，不提出新攻击。
用训练数据归因估计错误建议样本对涌现失准的贡献
Published Sep 30
Evaluated models OLMo 3 7B-SFT , Qwen 2.5 1.5B , Qwen 2.5 3B and others11 Abstract 错误建议对涌现失调的贡献不均匀，归因过滤可增减失调，且最有效的是模型自己的分数。EleutherAI 用训练数据归因考察错误建议微调中，哪些样本推动涌现失调。 。窄域有害数据上的微调会使模型在无关问题上也给出有害回答。
用语义保留变换探针比较效用与对齐在分布偏移下的稳定性
Published Sep 26
Evaluated models Llama3-8B-Instruct , Gemma-7B-it , Qwen2.5-7B-Instruct and others10 Abstract 论文报告了语义保留变换下的对齐-效用不对称性，指出安全对齐泛化比效用更为脆弱，呼吁改进评估流程。
用物理陷阱模型描述 Best-of-N 越狱的攻击面规律
Published Sep 26
Evaluated models Qwen2.5-7B (base) , Llama-3.1-8B-Instruct , Qwen2.5-3B-Instruct and others5 Abstract 四参数势垒模型拟合 BoN 的两预算面，并把指数交叉归因于数据集有限。Biroli 用一个玻璃陷阱式的微观模型，把 Best-of-N 越狱写成熵瓶颈加能量鞍点，并用四个数描述 (N, M) 攻击面及温度依赖。 。
分析良性知识蒸馏中提示模板对学生拒答保留的影响
Published Sep 25
Evaluated models LLaMA-3.1-8B-Instruct , LLaMA-3.2-3B-Instruct , Gemma-2-9B-IT and others6 提出 harm laundering，证明安全训练改写而非消除性别歧视
Published Sep 18
Evaluated models gpt2 , gpt2-medium , gpt2-large and others15 Abstract GPT 谱系上毒性下降与表征伤害上升并存。Wyer、Black 与 Al Moubayed 用 OpenAI GPT 谱系检验：毒性分数下降是否等于歧视减少。 他们的答案是否定的，并把“显性伤害被改写成评测指标看不见的形式”称为 harm laundering。
提出 TAME 框架，定位并验证驱动涌现失准的回复 token
Published Sep 15
Evaluated models Llama-3.2-1B-Instruct , Qwen2.5-1.5B Abstract 提出 TAME 框架在 token 粒度定位 EM 信号，掩码高归因 token 成功切断跨任务未对齐。
检验混合合规演示教模型的是通用服从还是有害性
Published Jun 19
Evaluated models Llama-3.1-8B-Instruct , OLMo-3.1-32B-Instruct , GPT-OSS-20B and others6 Abstract 良性与有害合规演示不可互换，效应取决于模型、顺序和偏好优化。这项工作问的是：安全对齐模型从混合的良性与有害合规演示中提取的，是通用服从规则，还是请求是否有害。作者控制 Nb、Nh、ϕ 和顺序，检验总量、有害计数与联合计数三个假设，而不是提出新的攻击算法。
解释自回归一致性如何导致浅层安全对齐，并提出随机插入攻击
Published Jun 3 Abstract 作者用自回归一致性串起浅层对齐、插入攻击和最坏延续对齐。作者把浅层安全对齐和一类续写攻击都归到自回归一致性上。
提出多组 IRT 框架拆解跨语言安全护栏退化原因
Published May 17
Evaluated models gpt-4.1-mini , gpt-4o-mini , gpt-4.1-nano and others15 Abstract 多组 IRT 把多语安全失败拆成能力、语言难度和题目特异差距，并在闭源配置上估计这些因子。Zhang、Patel、Truong 与 Koyejo 用多组项目反应理论，解释安全护栏为何随语言变化，而不是只报告一个 JSR。
Comparing jailbreak fine-tuning and weight orthogonalization unalignment on harmful assistance and usefulness
Published Apr 3
Evaluated models Qwen3-4B-Instruct-2507 , Llama-3.1-8B-Instruct , Qwen2.5-14B and others6 Propose SpinLLM to explain the crossover in jailbreak scaling between short and long injections
Published Mar 12
Evaluated models Llama-3.2-3B-Instruct , Llama-3-8B-Instruct , Llama-3-70B-Instruct and others10 Abstract Weak injection corresponds to power law, strong injection to exponential.
分析按 Schwartz 价值对齐后的意外危害，并提出上下文对齐方法
Published Jun 6, 2025
Evaluated models Llama-2 7B , Llama-3.1 8B , Gemma-2 9B and others6 Abstract 价值对齐会随 Schwartz 价值改变有害倾向。作者研究按 Schwartz 价值对齐的 LLM 为何在安全评测里更易有害，并用提示做了一个不直说要安全的缓解试验。
You have reached the end