Translation in progress Some entries still include text in the original language.
Oct 12 · Monday
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 ParanoiaEval 评测编码智能体的不必要风险处置
Published Oct 7
Evaluated models Opus-5 , Sonnet-5 , Sonnet-4.6 and others8 Abstract 作者用证据受控任务对评测风险处置，称其普遍、独立于任务能力并降低满意度。作者提出 ParanoiaEval，并称它是编码智能体风险处置能力的首个统一基准。要解决的问题是：证据已经确定风险应如何处置之后，智能体仍可能扩大验证、增加保护或备份，而现有评测要么只看功能正确性，要么把相关行为拆开研究。做法是把 NIST 的 ATMA 落到仓库任务。
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。
测量面向开发者的 MCP 错误信息对工具调用 Agent 恢复的影响
Published Sep 28
Evaluated models GPT-5.5 , GPT-5.6 Sol , GPT-6 Sol and others5 Abstract 面向开发者的 MCP 错误步骤会让只能调用工具的智能体停手。MCP 错误信息里的下一步常常按人类开发者来写，而只能调用服务器工具的智能体会照着做。做不到就结束回合。作者先统计这类文本，再在 BFCL V4 上只更换错误文本，看回合状态能否回到参考调用之后。
发现暴露模型家族标签会让多智能体派系化并削弱合作
Published Sep 28
Evaluated models Gemma (gemma-4-26B-A4B-it) , GPT-OSS (gpt-oss-20b) , Nemotron (NVIDIA-Nemotron-3-Nano-30B-A3B-BF16) and others6 Abstract 可见身份标签让异构 LLM 群体按标签结派，共识更慢、更贵，去掉标签后该行为消失。Del Giudice 等人研究异构多智能体 LLM 系统里一个配置细节：把模型家族身份告诉智能体，会不会改变合作。他们把由此出现的、按身份结组的倾向称为 factionalism 。任务是无利害的共同决策，没有人设，也没有偏向同家族同伴的奖励。
提出 BioSecBench-Refusal，评测智能体生物研发拒答与隐蔽危害识别
Published Jul 6
Evaluated models Opus 4.8 , Opus 4.7 , Opus 4.6 and others10 Abstract 配对基准显示拒绝多由 API 触发，合法任务常被拒得不少于隐蔽危害。BioSecBench-Refusal 把生物研发智能体的谨慎程度和识别隐蔽危害的能力放在同一套任务上测量。
提出 RefusalBench，评测生物研究提示上的拒答率与安全校准
Published May 20
Evaluated models Claude Opus 4.7 , Claude Opus 4.6 , Claude Opus 4.5 and others15 构建 IndicSafe 基准，评测 Indic 语言安全标签的跨语言一致性
Published Mar 19
Evaluated models GPT-4o Mini , Claude Sonnet v4 , Grok-3 and others10 测量任务无关说服对 Agent 编码与网页研究行为的影响
Published Feb 1
Evaluated models gpt-4.1-nano , mistral-nemo-12b , llama-3.1-8b Abstract 任务无关说服会在轨迹上留下任务相关的行为痕迹，但立场易感性不能稳定预测其幅度。这篇工作测量一件事：与任务无关的说服进入对话之后，AI 智能体执行后面的编码或网页研究时，轨迹是否系统性不同，以及“更容易被说服”能否预测差有多大。
You have reached the end