MIA-DPO: Multi-Image Augmented Direct Preference Optimization For Large Vision-Language Models Paper • 2410.17637 • Published Oct 23, 2024 • 34