Translation in progress Some entries still include text in the original language.
Oct 12 · Monday
All papers Found 70 papers Another 233 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.
Trace attention to identify harmfulness to the MLP refusal write pathway and propose Sentinel
Published Oct 9
Evaluated models Llama-3.1-8B-Instruct , Qwen2.5-7B-Instruct , Qwen3-8B-Instruct and others18 Abstract Attention updates are more readable for harmfulness, MLP updates write refusal, and Sentinel routes alerts using two-path signals.The authors reframe jailbreaking and excessive refusal of aligned models as a pathway problem on the residual stream from recognition to writing, and accordingly provide conditional reminders before decoding.
Uses premise permutation and hidden-layer tracing to audit whether legal chain-of-thought truly relies on the cited legal provisions.
Published Oct 9
Evaluated models Qwen3-8B , Qwen3-14B , Llama-3.1-8B and others7 Abstract Using counterfactual substitution of authorities and internal commitment tracing, the authors find that legal CoT rulings depend on authorities only to a limited extent and are vulnerable to prompt injection manipulation.This study conducts an empirical audit of the causal truthfulness of legal citations in chain-of-thought reasoning of large language models in the legal domain.
提出零空间激活引导 Nullify，免训练擦除指定记忆并保持效用
Published Oct 8
Evaluated models Llama-2-7B-chat-hf , Qwen3-8B 提出 PatchBench-Local 协议，评测激活补丁修复越狱的局部附带损伤
Published Oct 7
Evaluated models Qwen2.5-3B-Instruct , Qwen2.5-14B-Instruct , gemma-3-4b-it and others8 Abstract 论文提出 PatchBench 与局部评估协议，揭示了激活补丁在全局能力无损下仍存局部良性退化。
提出 Safety-Aware Pruning，缓解视觉 token 剪枝放大的多模态越狱
Published Oct 7
Evaluated models LLaVA-1.5-7B Abstract 剪枝去掉背景 token 后注意力塌向恶意前景。这篇工作评估视觉 token 剪枝如何改变 VLM 在多模态越狱下的行为，并给出推理时防御 SAP。
提出 Persist-3 序贯检验，从隐藏表示识别未见越狱源
Published Oct 7
Evaluated models Llama2-7B-chat , Qwen2.5-7B-Instruct , Gemma-2-9B-it and others4 Abstract Persist-3 用有限窗口检验攻击流是否超出已知越狱源，检出率低于全历史 SR-DKW，但决策只依赖最近观测。Persist-3 是一个有限记忆的序贯检测器，用来判断攻击表示流是否已经不能由任何一个已知越狱源解释。 。
测量有损投机解码的安全代价，并提出 SecureSD 收紧早期验证
Published Oct 7
Evaluated models Qwen3-8B , Qwen3-32B , Llama-3.1-8B-Instruct and others6 Abstract 有损推测解码的早期偏离更快抬高越狱与注入风险。作者研究有损推测解码在换取速度时，是否会比效用更快地损害越狱与提示注入防护，并给出按位置校正的 SecureSD。
提出 MARCO 放射性水印，冻结蛋白质生成模型并在去噪时嵌入签名
Published Oct 6
Evaluated models RFdiffusion , ESMFold , Chroma and others6 Abstract MARCO 在冻结的 PGM 上去噪时嵌构象水印，作者称能兼顾保真与抗扰动，并使盗版模型继承水印。MARCO 是加在蛋白质生成模型推理过程上的构象水印，用来标记知识产权，并为可能的生物安全滥用保留可追溯信号。作者还把它说成数字权利管理工具，以及符合生物武器公约的数字监管链。问题在于：专有 PGM 的三维输出可能被拿去训练盗版模型，也可能在来源被抹掉后绕过筛查。
提出 PEV 攻击，在输入嵌入上加高斯噪声越狱开源模型
Published Oct 6
Evaluated models Llama-3.1-8B-Instruct , Qwen2.5-1.5B-Instruct , Phi-4-mini-3.8B and others6 Abstract PEV 用 embedding 高斯噪声，在六个模型上越狱 JailbreakBench 全部提示。PEV 是对 open-weight LLM 的 embedding 噪声越狱。
提出 ReSA，从对齐模型行为恢复安全奖励并在解码时反转
Published Oct 6
Evaluated models Llama3.1-8B-Instruct , Gemma-7B-Instruct , Qwen2.5-7B-Instruct and others9 Abstract ReSA 用最大熵 IRL 从对齐模型行为恢复代理安全奖励，并在解码时反转以诱导有害输出。ReSA 是一种对抗攻击：从对齐模型的可观察行为恢复代理安全奖励，再在解码时把该奖励反过来用。 。
提出 TRACER，从权重差识别扩散模型被擦除概念
Published Oct 5
Evaluated models Stable Diffusion v1.5 , Stable Diffusion 3 , FLUX.1 and others5 Abstract TRACER 用权重足迹识别未披露擦除概念并估计个数，作者称数秒内完成且准确率高于对照。TRACER 针对的是遗忘扩散模型上的目标识别：提供者从公开基座 W 用已发表方法擦掉未知集合 S⋆，只发布 W′。攻击者还有假定包含这些概念的词表，但不知擦了谁、擦了几个、用了什么算法。
提出生成时防御 SENTINEL，匹配输入输出以抽出越狱意图并停写
Published Oct 3
Evaluated models Llama2-7B , Llama3-8B , Mistral-7B-v0.2 and others6 提出 UnAct，仅以前向激活干预从分类器中遗忘指定类别
Published Oct 3
Evaluated models ResNet-18 , ViT-B/16 Abstract UnAct 用前向激活做类别遗忘。UnAct 是一种事后类别遗忘：删除请求可以只有少量该类图像，且没有标签和保留数据。它不用梯度，只对遗忘图像做前向。编辑对象是分类器所读的后期单元。
提出基准 OLMo-Detect 评测 LLM 成员推断
Published Oct 2
Evaluated models OLMo 2 1B , OLMo 2 7B , OLMo 2 13B and others9 用分类后缀提升激活探针对分布外恶意输入的检测
Published Oct 2
Evaluated models Llama-3.1-8B , Qwen3.5-9B , Gemma-4-12B Abstract 作者称排序增益来自分类格式，低 FPR 精度来自点名准则。用户轮后的分类后缀，被作者用作激活探针的分布外检测干预：排序看格式，低 FPR 精度看是否点名恶意/良性。
用因果激活修补定位模型服从提示注入的决策层
Published Sep 29
Evaluated models Qwen3-4B , Qwen3-14B , Qwen3-32B and others7 Abstract 合规的因果杠杆集中在网络后三分之一的低秩子空间，该层也是混淆下检测最好的位置。这篇工作用逐层激活修补定位提示注入合规发生在网络何处，并把该位置用于检测验证。 。
比较用于审计提示注入与隐瞒的 token 位置选择信号
Published Sep 29
Evaluated models Qwen2.5-7B , Gemma-3-12B , Gemma-3-27B and others4 提出用扰动蒸馏本地代理并离线过滤来提取黑盒模型凭据
Published Sep 29
Evaluated models DeepSeek-Coder-7B , LLaMA-3-8B , Qwen2.5-Coder-7B and others6 Abstract 论文提出扰动蒸馏与离线过滤两阶段框架，实现了 output-only 访问下延迟降低的黑盒大模型 API key 提取。
提出 Constitutional adapters，蒸馏宪法行为以缓解越狱与失准
Published Sep 29
Evaluated models Qwen3.5-4B , Llama-3.1-8B-Instruct , Gemma-4-E4B and others4 Abstract 差分宪法适配器在长上下文和多轮上更稳，并可按 α 换取遵从。
诊断护栏相对基座在对抗提示上的置信失校
Published Sep 29
Evaluated models Llama-Guard 7B , Llama-Guard-2 8B , Llama-Guard-3 8B and others10 Abstract 对抗包装使 guard 高置信漏检，而对应 base 在同一输入上常更不确定。诊断性分析 ：五个提示词分类 guard 的置信度，在清洁输入上可以较校准，在越狱包装后则变成高置信度的错误判定。