Zero-shot learning

Definition

Zero-shot learning allows large language models to perform tasks without any task-specific examples in the prompt, relying solely on instructions and pre-trained knowledge to generalize to unseen tasks. It tests the model's ability to understand and apply concepts from training data alone.

Models Using Zero-shot learning(3)