1 paper
Peng Kuang, Xiangxiang Wang, Wentao Liu +2
Multimodal Large Language Models (MLLMs) have achieved impressive performances in mathematical reasoning, yet they remain vulnerable to visual hallucinations and logical inconsiste…