TY  - RPRT
TI  - TAB: Transformer Attention Bottlenecks enable User Intervention and Debugging in Vision-Language Models
AU  - Pooyan Rahmanzadehgervi
AU  - Hung Huy Nguyen
AU  - Rosanne Liu
AU  - Long Mai
AU  - Anh Totti Nguyen
PY  - 2025
UR  - https://arxiv.org/abs/2412.18675
ID  - 2412.18675
ER  -