1 paper · 2 filters
Shuhang Chen, Hangjie Yuan, Yunqiu Xu +5
Despite strong results on many tasks, multimodal large language models (MLLMs) still underperform on visual mathematical problem solving, especially in reliably perceiving and inte…