AI Glossary
In-Context Learning
AI adapting to tasks from examples in the prompt
Definition
In-context learning (ICL) is the ability of large language models to perform new tasks by learning from examples provided within the prompt itself — without updating the model's weights. When you give a model several question-answer examples and then ask it a new question, it is using in-context learning. This is a key emergent capability of large models that enables few-shot prompting.