Few-shot Learning

Techniques

Simple Definition

Providing a small number of input-output examples in the prompt to guide the AI's response format and style.

Full Explanation

Few-shot prompting dramatically improves performance on structured tasks by showing the model exactly what format you want. Instead of describing the task, you show 2-5 examples of desired input→output pairs. The model infers the pattern and applies it to new inputs. Particularly effective for classification, data extraction, and stylistic writing tasks.

Example

Providing 3 example customer reviews with sentiment labels before asking the model to classify new reviews.

Related Terms

Zero-shot Learning

Asking an AI to perform a task it wasn't explicitly trained on, relying solely on its general capabilities.

Prompt Engineering

The practice of crafting inputs to AI models to reliably get better, more accurate, or more specific outputs.

Last verified: 2026-03-30← Back to Glossary