Super HN

   Cache-to-Cache: Direct Semantic Communication Between Large Language Models (arxiv.org) 3 points by jonbaer 3 hours ago