Skip to content
ai

Few-shot learning

Few-shot Learning

Definition

Few-shot learning is the ability of a language model to perform a new task by conditioning on a small number of input-output examples provided in the prompt context, without updating model weights. GPT-3 demonstrated that sufficiently large models can generalize from as few as 3–10 examples.

Few-shot prompting is a key technique for adapting LLMs to specialized tasks without fine-tuning.

Can this ship? Are we exposed? Can you prove it?

Every CISO is asking these three questions. We answer them with facts, not guesses, from repository to runtime.