Skip to main content
Artificial Intelligence

Weak-to-strong generalization

OpenAI Blog · · 1 views

We present a new research direction for superalignment, together with promising initial results: can we leverage the generalization properties of deep learning to control strong models with weak supervisors?

We present a new research direction for superalignment, together with promising initial results: can we leverage the generalization properties of deep learning to control strong models with weak supervisors?