Seeing Right but Saying Wrong: Inter- and Intra-Layer Refinement in MLLMs without Training

Add code
Jan 12, 2026

Share this with someone who'll enjoy it:

View paper onarxiv icon

Share this with someone who'll enjoy it: