Revisit What You See: Disclose Language Prior in Vision Tokens for LVLM Decoding

Open in new window