Search papers, labs, and topics across Lattice.
1
0
3
Forget expensive human feedback loops: a VLM-powered reward function can efficiently align image editing diffusion models with human preferences.
