Search papers, labs, and topics across Lattice.
1
0
2
Stateful Visual Encoders enable VLMs to leverage prior visual context, leading to substantial performance gains in multi-image tasks.
