1 paper
Jiachen Guo, Xinshan Zhu, Boyun Wang +4
Multimodal large language models (MLLMs) have made rapid progress in general visual understanding, yet they still struggle with fine-grained instance-level consistency verification…