"Designing agent incentives to avoid reward tampering", DeepMind

by gwern14th Aug 201915 comments

29

Ω 7

Outer AlignmentGoodhart's LawMachine Learning
Frontpage