In-Context Learning
The ability of a language model to learn a new task simply from instructions and examples provided within its prompt, without updating its weights.
Think of It Like This
Like showing a chef three examples of how you want your sandwich cut, and they immediately understand the pattern.
This emergent capability relies on the model's massive pre-training, allowing it to recognize patterns and adapt its output dynamically at inference time. It eliminates the need for expensive fine-tuning for many specialized tasks. However, its effectiveness is strictly limited by the model's context window size and prompt formulation.