Concept
Zero-shot
Solving a task with no labelled examples, relying only on a natural-language instruction.
Definition
Zero-shot prompting asks a model to perform a task it has never seen labelled examples for, using only a description. Frontier LLMs are surprisingly capable zero-shot across classification, translation and reasoning, though few-shot examples typically lift accuracy further.
Common use cases
- New domains
- Rapid prototyping
- Open-set classification