Generate, but Verify: Reducing Hallucination in Vision-Language Models with Retrospective Resampling

Published in NeurIPS 2025, 2025