What is the difference between zero-shot, one-shot, and few-shot prompting?

A They differ in model size requirements B Zero-shot: no examples in prompt. One-shot: one example. Few-shot: several examples. More examples generally improve performance by demonstrating the task format and expected output style C These terms describe the number of training runs D Zero-shot uses no model knowledge; few-shot uses training data