Cache-to-Cache: Direct Semantic Communication Between Large Language Models

13 points | by jonbaer 3 days ago