One-shot
Showing the model exactly one example before asking it to do the same thing.
Think of it like
Like a single “here’s how” demo before someone tries it themselves.
Example
You give one sample of the JSON shape you want, then the real input — and the model mirrors the structure.
How it actually works
One-shot sits between zero-shot and few-shot: a single demonstration is often enough to lock in a format or convention while spending almost no extra tokens. When one example carries the whole pattern, it’s the efficient sweet spot.
For product teams
A great token-cheap way to pin down output format without a full example set.
For engineers
In-context learning with a single demonstration; efficient when one example conveys the target format.
Related
Read anything AI without the jargon
Look up any term in plain English, or save terms as you read with the free Chrome extension.
Open DecoderAdd to Chrome