Search papers, labs, and topics across Lattice.
1
0
2
7
Scaling linear attention models with Sparse Delta Memory leads to significant improvements in long-context recall and reasoning without the computational burden of larger state sizes.
