Inverse scaling prize for finding a task where larger language models do worsegithub.com3 points·echen··0 commentsOpen articleSaveView on HN