Learning from Next-Frame Prediction: Autoregressive Video Modeling Encodes Effective Representations Paper • 2512.21004 • Published 1 day ago • 7
