Reinforcement learning in language models recruits a functional welfare axisfunctionalwelfare.com2 points·paraschopra··0 commentsOpen articleSaveView on HN