Towards Efficient and General-Purpose Few-Shot Misclassification Detection for Vision-Language Models

Open in new window