Paper

VIGOR: Benchmarking Visual Rationale Correctness in Multimodal Large Language Models · ByteTrail