Corrigibility as a Singular Target: A Vision for Inherently Reliable Foundation Models

LRMRe-assign community