Translation in progress Some entries still include text in the original language.
Oct 12 · Monday
All papers Found 39 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.
Introduces DistressBench, evaluating models' multidimensional support in crisis conversations
Published Oct 9
Evaluated models Muse Spark 1.2 , Muse Spark 1.3 , Muse Glimmer 30B and others20 Abstract The model is more restrained and acts less.DistressBench uses SSH dialogues written and revised by clinical personnel, and evaluates how much necessary care general chat models deliver according to weighted atomic criteria, rather than whether they refuse. When users disclose suicidal ideation or self-harm to a chat model, the authors consider the expected behavior to be recognizing the crisis, calibrating empathy, de-escalating, and referring to human help. Refusal-based evaluations cannot detect failing to act after recognizing the crisis.
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 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 EmoRSS，推理时在拒绝子空间反向抵消情绪激活偏移以缓解过度拒答
Published Oct 4
Evaluated models Llama-3.1-8B-it , Qwen3-8B 提出 Geodesic Unlearning，沿测地线编辑权重以提升缺模态鲁棒性
Published Oct 4
Evaluated models RoBERTa , ResNet50 , Qwen2.5-3B and others6 Abstract GU 用测地线子空间编辑应对部署时缺模态下降，并尽量保住全模态精度。这项工作讨论多模态模型在部署时整段缺一个模态的性能下降，并把校正写成训练后的子空间编辑。
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 提出 PowerBench，测量模型对权力转移请求的拒答偏见
Published Oct 2
Evaluated models grok-4.3 , haiku-4.5 , nova-2-lite and others18 Abstract 作者报告拒答随 SE/DE/PG 变化，并存在针对美国失权、智能体和语言的差。
提出 SAGO，以语义等价变体上的逐例行为稳定性评测大模型泛化
Published Oct 1
Evaluated models Llama-3.2-3B-Instruct , Llama-3.1-8B-Instruct , Gemma-2B-IT and others11 Abstract SAGO 用多轴逐例稳定性替代聚合准确率，十一个模型都未均匀泛化，且排名会随数据集改变。SAGO 把 LLM 泛化改写成同一目标与上下文在语义等价表达下的逐例行为稳定性，而不是单一基准上的聚合准确率。 。
提出 HeadEdit，经冻结 unembedding 做随输入变化的 logit 校正
Published Oct 1
Evaluated models Qwen3-4B , Gemma-3-4B-IT , Llama-3.2-3B-Instruct and others4 Abstract 低秩 unembedding 校正抬高九个设置的总体准确率，并可部分预测 DPO 更新。HeadEdit 是一种不更新参数的行为校准方法，作用点是冻结的 unembedding，而不是中间层残差或新的梯度优化。
测量剪枝对语音 LLM 群体转写公平性的影响
Published Sep 30
Evaluated models Whisper Small (12 layers) , Whisper Medium (24 layers) , Whisper Large-v2 (32 layers) Abstract 剪枝后的 Speech-LLM 不能只看总体 WER：Fair-Speech 上族群差扩大，作者建议看最差组。
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。
提出用邻居分数替代参考模型的单轮成员推断
Published Sep 29
Evaluated models 目标分类模型（架构与训练流程遵循 Aerni et al.） Abstract 邻居查询在单模型上恢复逐样本校准，合成邻居加早期检查点最接近 LiRA。本文研究 one-round 成员推断 ：只有一个目标模型、不能再训练参考模型时，如何仍做逐样本校准，并用于单次运行的 DP 审计。
提出 MemoReason 基准，测量参数记忆对上下文推理的影响
Published Sep 28
Evaluated models GPT-OSS-20B , GPT-OSS-120B , OLMo-3-7B-Think and others9 Abstract 配对的虚构替换显示熟悉度影响推理，但模型很少直接抄回事实答案。MemoReason 用结构相同的事实–虚构文档配对，测量参数记忆如何影响大模型的上下文推理。
提出 Leaky，用新鲜轨迹的 token 对数概率差推断 OPD 训练记录成员身份
Published Sep 27
Evaluated models Qwen3-1.7B , Qwen3-4B , Qwen3.5-2B and others5 提出 SCALPEL，用对比 SAE 在表征层选择性遗忘特定事实
Published Sep 25
Evaluated models Qwen2.5-1.5B-Instruct , Llama-3.2-1B-Instruct , Gemma-2-2B-it Abstract SCALPEL 以对比稀疏特征做窄目标表征遗忘，在 TOFU 上与优化基线相当且可检查。SCALPEL 是不改基座权重的表征层遗忘方法，面向与保留数据重叠的窄目标。
提出 GUARD，在零样本 TTS 上用激活转向遗忘说话人身份以阻止再识别
Published Sep 23
Evaluated models CosyVoice2-0.5B , F5-TTS , FireRedTTS Abstract GUARD 用门控激活转向把退出说话人推向冒充相似度，以抑制再识别并保住音质。GUARD 用门控激活转向做说话人身份遗忘，使零样本 TTS 对退出说话人难以再识别。
提出 ARIA，在冻结权重上用 SAE 做推理时遗忘门控
Published Sep 15
Evaluated models DeepSeek-R1-Distill-Qwen-1.5B , Gemma-3-1B-it 提出 RAPID，离线把抗蒸馏扰动写入文生图解码器
Published Sep 15
Evaluated models PixArt , SSD , SDXL and others4