Search papers, labs, and topics across Lattice.
1
0
2
Forgetting individual information in MLLMs can be achieved without compromising model performance, thanks to a novel method that localizes updates to specific layers.
