Skip to content
AImpact
IT EN
Inference Beginner Also known as: Apprendimento con pochi esempi

Few-shot learning

A prompting technique where the model is shown a few examples of desired input and output, so it learns the format on the fly without training.

ShareLinkedInX

In practice

Useful to enforce a schema, a tone, or a precise classification. Usually 3-5 examples are enough. It is almost always the first thing to try before reaching for fine-tuning: it only costs a few extra tokens in the prompt.

Related terms

Zero-shot learning Chain-of-thought

Seen in the wild

3 entries mentioning it
  1. IDEFICS: the first open-source replica of Flamingo
    High
  2. DeepMind Flamingo: the first few-shot visual language model
    High
  3. GPT-3: the paper that opens the scaling-laws era
    Landmark
← All terms