RoboHarm: Do Frontier Robot Policies Refuse Unsafe Instructions?

发生了什么

Three robot policies given five instructions they should refuse. How often they refused, how often they carried them out.

摘要按规则整理自下方来源原文

时间线

1. 18:58
   Hacker News 最先出现HHacker News
2. 11:59
   Lemmy Technology 出现讨论LLemmy Technology

来源