LongVPO: From Anchored Cues to Self-Reasoning for Long-Form Video Preference Optimization Paper • 2602.02341 • Published Feb 2 • 1
