Translation in progress Some entries still include text in the original language.
Oct 12 · Monday
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 KANSteer，用非线性激活引导拟合文生图有序概念轨迹
Published Oct 3
Evaluated models FLUX.1-schnell , FLUX.1-dev Abstract KANSteer 用曲线替代固定方向。KANSteer 研究文生图扩散模型里，有序视觉概念能否沿单一激活方向被遍历，并改用一维曲线做 steering。
提出 D-Scope，用视觉质心检索 SAE 解码方向并转向扩散生成
Published Sep 30
Evaluated models SANA-Sprint , Nitro-1-PixArt , SANA teacher and others4 Abstract D-Scope 用视觉质心检索单个 SAE 方向做掩码转向，对比检索提高区域增益但不一致改善保持。D-Scope 把扩散 Transformer 中 SAE 特征的高激活图像块，接到文本查询驱动的单方向生成干预上。
用逐步 SAE 诊断文生图概念脆性，并推理时向类原型插值
Published Sep 9
Evaluated models SD 1.5 , SD 3.5 , SDXL and others5 Abstract 逐步 SAE 诊断物体依赖的概念脆性，早期向类原型插值可在五骨干上提高风格与属性指标。作者用逐步稀疏自编码器审计文生图扩散模型里随物体变化的概念脆性，并用推理时原型插值检验该诊断。 。
提出 KPI，用入度定位跨层关键路径并做 SAE 转向以解决知识冲突
Published Sep 8
Evaluated models Llama-3-8B , Llama-3.1-8B , Gemma-2-9B Abstract KPI 以入度关键路径替代大批 SAE 特征转向，在冲突 RAG 上提高忠实并减轻副作用。KPI 针对知识冲突 RAG 中的 SAE 转向：相关性批量改特征会带入冗余，少量因果上关键的特征即可完成转向。
用 SAE 追踪稠密检索查询扩展的潜变量并用激活转向验证
Published Sep 7
Evaluated models Qwen3-Embedding-0.6B（Qwen-Embed） , SimLM Abstract SAE 追踪稠密检索 QE 的潜变量，转向在四基准上比文本扩展更稳。作者用稀疏自编码器分析稠密检索里的查询扩展：现成检索器不微调时，拼接扩展文本往往增益有限甚至变差，而最终指标和嵌入整体偏移不足以说明内部发生了什么。
You have reached the end