Skip to content
ai

In-context learning

In-context Learning

Definition

In-context learning (ICL) is the ability of large language models to adapt their behavior based on examples or instructions provided within the prompt, without gradient updates to model weights. The model implicitly "learns" from the examples at inference time.

ICL is the mechanism behind few-shot, one-shot, and zero-shot prompting and is an emergent property of large-scale pre-training.

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.