Translation in progress Some entries still include text in the original language.

METR · News

AI Safety HOTOct 13 · Tuesday

arXivY

METR · News

Latest publications and developments

Latest publications and developments

Search titles, summaries and full text/

METR

Research topics

Prompt injectionAgent safetyDeception and schemingReward hackingChain-of-thought monitoringInterpretabilityAI controlBackdoors and poisoning隐私与数据泄露Supply-chain securityDangerous capabilitiesAI 与网络攻防开源模型安全Red teaming

Companies and institutions

OpenAIAnthropicGoogle DeepMindGoogle ResearchGoogle 安全MetaxAIMicrosoftNVIDIADeepSeek阿里巴巴 · Qwen智谱 · GLM月之暗面 · Kimi字节跳动腾讯蚂蚁集团上海人工智能实验室UK AISIUS CAISIMETRApollo ResearchRedwood ResearchFAR.AIPalisade Research

Reports and frameworks

System CardFrontier safety frameworks

TimelineTrendingTodaySelected

METR · News

Latest publications and developments

TimelineTrendingTodaySelected

METR

Research topics

Prompt injectionAgent safetyDeception and schemingReward hackingChain-of-thought monitoringInterpretabilityAI controlBackdoors and poisoning隐私与数据泄露Supply-chain securityDangerous capabilitiesAI 与网络攻防开源模型安全Red teaming

Companies and institutions

OpenAIAnthropicGoogle DeepMindGoogle ResearchGoogle 安全MetaxAIMicrosoftNVIDIADeepSeek阿里巴巴 · Qwen智谱 · GLM月之暗面 · Kimi字节跳动腾讯蚂蚁集团上海人工智能实验室UK AISIUS CAISIMETRApollo ResearchRedwood ResearchFAR.AIPalisade Research

Reports and frameworks

System CardFrontier safety frameworks

Sep 27Sun

Sep 27Sunday · 1 items

1. 15:00
   METR · Sep 27 15:00Attention4646

   METR 发布逐动作监控研究笔记：为自家评测部署实时阻断监控器

   METR 研究人员发布研究笔记，介绍其为自家 Agent 评测部署的逐动作实时监控器：由 LLM 裁判在每个工具调用执行前打分，超过阈值即暂停评测并转人工复核。

You have reached the end