interesting research from OpenAI on how AI models are getting more difficult to monitor through CoT, laying out a few reasons. One finding from the blog:
"... we expect further progress in pretraining to require new approaches to monitorability."
OpenAI research finds AI models are harder to monitor through their chain of thought
AIOpenAI shares research on how chain-of-thought monitorability, covering no-CoT capability and CoT controllability, changes over the model training lifecycle. The research finds models become harder to monitor through their chain of thought as they scale toward more intelligent systems.
Max ZeffVerified on X
@ZeffMax
(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
Max Zeff · x.comPublished · added here
Original title: interesting research from OpenAI on how AI models are getting more difficult to monitor through CoT, laying out a few reasons. One findin...