A Unified Hallucination Mitigation Framework for Large Vision-Language Models

Add code
Sep 24, 2024

Share this with someone who'll enjoy it:

View paper onarxiv icon

Share this with someone who'll enjoy it: