Towards Mitigating Hallucinations in Large Vision-Language Models by Refining Textual Embeddings

Aakriti Agrawal, Gouthaman KV, Rohith Aralikatti +6

Hallucinations in Large Vision-Language Models (LVLMs) remain a persistent challenge, often stemming from inadequate integration of visual information during multimodal reasoning.…