1 paper · 1 filter
Jiarui Liu, Weihao Xuan, Zhijing Jin +1
Multimodal Large Language Models (MLLMs) often suffer from hallucinations, particularly errors in object existence, attributes, or relations, which undermine their reliability. We…