TY  - RPRT
TI  - A Comprehensive Information-Decomposition Analysis of Large Vision-Language Models
AU  - Lixin Xiu
AU  - Xufang Luo
AU  - Hideki Nakayama
PY  - 2026
UR  - https://arxiv.org/abs/2603.29676
ID  - 2603.29676
ER  -