Seeing is Believing? Mitigating OCR Hallucinations in Multimodal Large Language Models

Open in new window