Enhancing Vision-Language Model Reliability with Uncertainty-Guided Dropout Decoding

Open in new window