Anthropic: Decomposing Language Models into Understandable Componentsanthropic.com4 points·wodow··0 commentsOpen articleSaveView on HN