Search papers, labs, and topics across Lattice.
University of California Riverside
1
0
3
6
Turn frozen vision-language models into off-the-shelf safety supervisors for text-to-image generation, sidestepping the need for fine-tuning or curated datasets.