🤖 Same Tokenization Pattern!

Camera images become ViT patch tokens

Joint angles become proprioceptive tokens

One unified sequence → Same transformer encoder!