1 paper
Songyuan Yang, Weijiang Yu, Jilin Ma +5
Video reasoning has advanced with large multimodal models (LMMs), yet their inference is often a single pass that returns an answer without verifying whether the reasoning is evide…