TY  - RPRT
TI  - VISOR++: Universal Visual Inputs based Steering for Large Vision Language Models
AU  - Ravikumar Balakrishnan
AU  - Mansi Phute
PY  - 2026
UR  - https://arxiv.org/abs/2509.25533
ID  - 2509.25533
ER  -