技术博客/SHADE-Arena: Evaluating sabotage and monitoring in LLM agents
Anthropic进阶2025-06-16· 11 分钟· Agent 智能体
SHADE-Arena: Evaluating sabotage and monitoring in LLM agents
Alignment Jun 16, 2025 Read the paper As AI models get smarter, we also need to become smarter in how we monitor them. More intelligent systems can ta
本文需解锁后阅读
免费开放 3 篇精选工程博客（列表置顶）。其余文章登录后可阅读全文。