VLA-JEPA: Enhancing Vision-Language-Action Model with Latent World Model Paper • 2602.10098 • Published 7 days ago • 16