Improving Zero-shot Generalization of Learned Prompts via Unsupervised Knowledge Distillation

Open in new window