Picture for Abdullah Ibne Hanif

Abdullah Ibne Hanif

Beyond Dominant Patches: Spatial Credit Redistribution For Grounded Vision-Language Models

Add code
Feb 25, 2026
Viaarxiv icon