Cache-to-Cache: Direct Semantic Communication Between Large Language Modelsarxiv.org14 points·jonbaer··0 commentsOpen articleSaveView on HN