Rethinking the Role of Efficient Attention in Hybrid Architectures Paper • 2606.15378 • Published 13 days ago • 17
The MiniMax-M2 Series: Mini Activations Unleashing Max Real-World Intelligence Paper • 2605.26494 • Published about 1 month ago • 41
Gated DeltaNet-2: Decoupling Erase and Write in Linear Attention Paper • 2605.22791 • Published May 21 • 33