Fetching the paper…

Unsupervised Visual Chain-of-Thought Reasoning via Preference Optimization · Around