elvis· @omarsar0 · X·· 19 小时前AI 评分58
AI 导读
elvis 推荐了 OpenAI 一项关于 CoT 可监控性的研究，该研究考察预训练、架构、强化学习和模型生成数据如何影响 no-CoT 能力与 CoT 可控性这两个代理指标。作者指出，模型越智能，通过 CoT 进行监控就越困难，许多问题仍无答案，但这项研究为探究该现象如何以及在何处出现提供了不错的框架。
正文 · 原文
Recommended read. They find that increasingly intelligent models are harder to monitor through their CoT. Lots of things remain unanswered, but this is a nice setup for studying how and where this emerges.
(1/10) Today, we’re sharing new research from @OpenAI: How does CoT monitorability (no-CoT capability and CoT controllability) evolve over the model training lifecycle? We emerged with unsurprising but sobering findings: as we scale toward more intelligent systems, models are becoming harder to monitor through their CoT. 🧵在 X 查看被引用的帖子
来源：elvis · x.com