#MLLMs

Woodpecker offers a novel training-free solution to the challenge of hallucinations in multimodal large language models (MLLMs). By employing five sequential stages: key concept extraction, question formulation, visual knowledge validation, visual claim generation, and hallucination correction, Woodpecker enhances MLLM accuracy by aligning text with corresponding images. This method is easy to implement, interpretable, and demonstrates marked improvements in model accuracy without requiring retraining, as evidenced by a 30.66% accuracy gain in evaluations.