Free preview5 min

What Is Few-Shot Prompting?

Reviewed by Human · Updated August 5, 2026

The myth: to get better AI output, you describe what you want more precisely. More adjectives, more rules, more edge-case caveats. The truth: past a certain point, description stops working — and the fix is to stop describing and start showing.

Why would three short examples beat three careful paragraphs of instructions?

What Is Few-Shot Prompting?

**Few-shot prompting**: including a small number of worked input-output examples inside your prompt so the model can copy the pattern. **Zero-shot prompting**: asking the model to do a task with instructions only — no examples. **In-context learning**: the model's ability to pick up a task pattern from examples in the prompt itself, without any retraining. This isn't a hack. It's the headline finding of the original GPT-3 paper, which showed that large models can learn tasks from demonstrations placed directly in the prompt (arxiv.org/abs/2005.14165). The model treats everything you write as evidence about what should come next — which is also why context is everything when you build a prompt. Examples are simply the highest-density evidence you can provide, and they slot into the same skeleton you learned in the anatomy of a prompt: role, context, task, examples, format.

One to five examples in the prompt = few-shot. Zero examples = zero-shot. Exactly one = one-shot.

Live Demo — AI Terminal

Here's a support-ticket classifier. The zero-shot version — "Classify this ticket by category and urgency" — returns a different format every run: sometimes a sentence, sometimes bullet points, sometimes labels nobody defined. Now watch what three examples do. Run this and check the output format.

Your prompt

Your turn

Take this zero-shot prompt and convert it to few-shot. Write two example pairs that lock in the output format you want.

Reflect

Which formatting decisions did your examples make for you that the original instruction left ambiguous?

📚

Go deeper

end of module

You've finished this module.

Mark it complete to earn your XP and keep your streak alive.

Progress saved locally · Sign up to earn XP

Designing Good Examples→