NeurIPS2025 1 [Paper Review] Vision-Language-Vision Auto-Encoder: Scalable Knowledge Distillation from Diffusion Models Jan 14, 2026