Detecting and reducing scheming in AI models

公司动态 OpenAI 原文语言:英文

Apollo Research and OpenAI developed evaluations for hidden misalignment (“scheming”) and found behaviors consistent with scheming in controlled tests across frontier models.

原文摘录

Apollo Research and OpenAI developed evaluations for hidden misalignment (“scheming”) and found behaviors consistent with scheming in controlled tests across frontier models.

以下是来源提供的简短摘录,完整内容请查看原文。

参考来源

← 返回每日简报