Skip to main content
All terms
Foundations

One-Shot Learning

Learning to perform a task from a single example.

Definition

One-shot learning is the ability to perform a task from a single example. In the context of large language models, it usually means giving one demonstration in the prompt before the real input, which is a form of in-context learning. It sits between zero-shot, where no example is given, and few-shot, which provides several.