1 paper
Zhe Qian, Nianbing Su, Zhonghua Wang +6
Current multimodal models often suffer from shallow reasoning, leading to errors caused by incomplete or inconsistent thought processes. To address this limitation, we propose Self…