Search papers, labs, and topics across Lattice.
1
0
3
Forget rote memorization: this self-gated training method teaches models to actually *learn* new information by focusing on what surprises them most.