Multimodal Autoregressive Pre-training of Large Vision Encoders | IEEE Conference Publication | IEEE Xplore