Thinking Before Looking: Improving Multimodal LLM Reasoning via Mitigating Visual Hallucination

Open in new window