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.
OpenAI research finds smarter models are harder to monitor through CoT
AIOpenAI researchers study how chain-of-thought monitorability, covering no-CoT capability and CoT controllability, changes across the model training lifecycle. They report that as models become more intelligent, they are harder to monitor through their CoT.
elvisVerified on X
@omarsar0
(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. 🧵
Source details
elvis · x.comPublished · added here
Original title: Recommended read. They find that increasingly intelligent models are harder to monitor through their CoT. Lots of things remain unanswere...