MM-OPD: Towards One More Bottleneck Between Perception and Reasoning
Recent multimodal large language models (MLLMs) advance visual reasoning by strengthening both perception and reasoning, implicitly assuming a process that transitions seamlessly from perception to reasoning. However, we observe a counterintuitive phenomenon that challenges this assumption: holding the model, question,...
Jin-Tao Tong, Yujing Lou, Zhan-Ming Shen et al.
· 0 citations