Alex Mallen
@alextmallen
Redwood Research (@redwood_ai)
Prev. @AiEleuther
参加 August 2021
344 フォロー中 1.1K ファン
@TheStalwart These are different concepts. Overfitting is an issue that happens with perfect training data. Reward hacking happens when training data is imperfect.
人気のあるユーザー
