In-Context Learning

The ability of large language models to adapt their behavior based on examples provided in the prompt, without any parameter updates or fine-tuning. In-context learning enables models to perform specialized tasks by simply showing them examples of desired input-output pairs during inference.