https://www.lesswrong.com/posts/pEZoTSCxHY3mfPbHu/catastrophic-goodhart-in-rl-with-kl-penalty (AI safety, decision theory, probability theory)
https://www.lesswrong.com/posts/pEZoTSCxHY3mfPbHu/catastrophic-goodhart-in-rl-with-kl-penalty
View External Link: