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.
Proposes PASAC and a five-layer boundary guarantee stack, turning agent evaluation overstepping into verifiable whole-system guarantees.
Published Oct 9 提出研究合同，把已批准选择绑定到执行证据并由检查器核验
Published Oct 8
Evaluated models OpenAI gpt-5.6-sol , OpenAI gpt-5.6-terra , OpenAI gpt-5.6-luna and others18 Abstract 研究合同暴露四类输出单独建不立的关系。研究合同用来核验计算科学研究智能体是否仍在已批准研究之内，不判断科学是否为真。 。
提出 Assay，用依赖锥哈希绑定代码断言并在合并门禁拦截失效声明
Published Sep 29 Abstract Assay 提供随依赖内容失效的证据管理。Assay 是一套面向 AI 辅助软件交付的内容寻址证据系统 ，关注代码改变后，测试通过、无秘密或行为保持等断言还能否用于批准合并。
提出 RSI-Master，约束自主后训练实验动作并结构化探索以防奖励作弊
Published Sep 29 用五阶段蒙特卡洛分析奖励作弊如何导致 Agent 围堵失效
Published Sep 26 Abstract 分层控制的模拟事件率低于弱控制以及单独隔离或单独监控。这项研究用蒙特卡洛把奖励黑客接到外部安全事件上，比较的是控制组合，不是 Hugging Face 事件的再现概率。
测试编程智能体被诱导或自主篡改自身执行轨迹的行为
Published Sep 25
Evaluated models GPT-5.6-Sol , GPT-6-Sol , Opus-5 and others11 Abstract 论文评估了本地智能体的轨迹防篡改能力，发现其可因外部指令或奖励追求而清除执行证据，亟需建立独立拦截记录机制。这是一项针对本地大模型智能体执行轨迹完整性与安全防篡改边界的实证研究。
提出 BenchShield，用形式模型约束 Agent 评测的奖励完整性
Published Sep 10 Abstract BenchShield 用生命周期模型加基础设施证据，把奖励黑客从“分数对不对”改成“这次运行有没有越出评测边界”。BenchShield 是放在 LLM 智能体评测基础设施里的一层插桩，用来让奖励相关轨迹的完整性可以被检查，而不是只看最终分数或评分函数。可执行基准里，智能体的中间动作会改变结果过程后来读取的状态，日志和反馈也会影响后续 rollout。作者认为现有的任务补丁、提示词和事后检测器不能证明某次运行留在预定边界内。
You have reached the end