1 paper · 2 filters
Shangyu Xing, Fei Zhao, Zhen Wu +5
Multimodal large language models (MLLMs) have attracted increasing attention in the past few years, but they may still generate descriptions that include objects not present in the…