Translation in progress Some entries still include text in the original language.
Oct 12 · Monday
All papers Found 12 papers Another 240 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.
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。
提出 Safety Operator，用谱优化调节安全指令表达强度
Published Sep 29
Evaluated models Gemma 3 1B , Gemma 3 4B 提出 RLCDAlignBench，零样本检测多种对齐失败
Published Sep 24
Evaluated models Jev (jev-1.13.0) Abstract 一次通用问题已能给多数对齐失败排序，大的误差来自状态和标签。RLCDAlignBench 测量用强化学习做校准决策的 Jev，能否当作对齐失败的零样本检测器。
揭示推理加深引发的对齐坍塌，用 Reasoning Trap 放大越狱
Published Sep 8
Evaluated models Qwen3-8B , Qwen3-14B , Llama2-7B and others8 Abstract 更深推理提升题目准确率，同时提高扰动下的相对损失。作者研究扩展推理是否以对齐稳健性为代价，提出 ALR、RT 与 RRA。 。
提出 TIER 基准，按威胁隐晦度评测有害提示下的安全行为
Published Sep 4
Evaluated models Qwen1.5-MoE , Gemma-2B , Mistral-7B and others6 在写入点定位拒答方向，经刚性旋转跨架构迁移并门控放大
Published Sep 4
Evaluated models Falcon-Mamba-7B , Falcon3-Mamba-7B , Llama-3.1-8B and others8 Abstract 拒答方向可经刚性旋转跨架构迁移，但必须在各架构的写入点读取。作者研究拒答表示是否随架构改变，结论是表示共享、读取位置随架构变。
提出 Safin-1，以可路由 Safety State 在冻结主干上实现内在安全
Published Aug 31
Evaluated models Safin-1 (4B) , Safin-1 (35B-A3B) 提出 aiXamine 黑盒基准，统一评测 LLM 安全、稳健与隐私权衡
Published Aug 21
Evaluated models openai/gpt-5-mini , openai/gpt-4o , anthropic/claude-haiku-4.5 and others15 Abstract aiXamine 用九服务黑盒协议比较百余个模型，作者称单轴信任度进展可能削弱其他轴。aiXamine 是一套黑盒评测协议，用来在同一条件下比较 LLM 的 safety、security 与 privacy，而不是再产出一个能力分。
提出 WIFA 意图组监督微调，缓解包装式越狱与过度拒答
Published Aug 13
Evaluated models Qwen2.5-7B-Instruct , Llama-3.1-8B-Instruct Abstract WIFA 用匹配意图组做数据层，WIFA-Boost 偏高安全，A-GCRT 在 Qwen 上把过拒降到基座以下。WIFA 是一种自动意图组数据构造，用来让安全微调拒绝有害意图而不是拒绝表面包装。WIFA-Boost 与 A-GCRT 在同一数据层上选取不同的安全–过拒工作点。
提出 CompositeHarm，比较句法操纵与语义危害的跨语言迁移
Published Feb 8
Evaluated models GPT-OSS 20B , LLaMA-3-8B-Instruct , Qwen3-32B 提出 CSSBench 评测轻量模型对中文表面对抗改写的失守与过度拒绝
Published Jan 2
Evaluated models Qwen3-0.6B , Qwen3-1.7B , Qwen3-8B and others10 Abstract CSSBench 用表面扰动和边界查询评轻量模型。CSSBench 评测轻量 LLM 在中文特有表面改写下的安全，并同时测量过度拒绝。要处理的缺口是：英文向护栏和只测干净查询的中文基准，看不清同音、拼音、符号混写、零宽字符这些写法下的失守，也分不清较低风险来自细粒度判断还是粗粒度拒答。作者称轻量模型更需要这种评测。
You have reached the end