MiniMax M2 Tech Blog 3: Why Did M2 End Up as a Full Attention Model?twitter.com1 point·logicprog··0 commentsOpen articleSaveView on HN