In-context learning is when a language model picks up how to do a task from examples or instructions in the prompt, without being retrained.
In plain words
You can show a model a few worked examples in your message and it follows the pattern for the next one. Nothing about the model changes: the learning happens only inside that one conversation. Researchers found that larger language models are much better at this.
A quick example
You paste three product reviews, each followed by "positive" or "negative", then add a fourth review with no label. The model answers "negative" in the same style, even though nobody trained it on your labels.
Where you'll meet it
- Few-shot prompts that include worked examples.
- Getting a chatbot to follow a format without fine-tuning.
Related
Sources
- Docsin-context learning, Google Machine Learning Glossary · read 29 Sept 2026
- PaperLanguage Models are Few-Shot Learners, Brown et al. (arXiv) · read 29 Sept 2026
Checked by WhatIsAI against the sources above · 29 Sept 2026. Spotted an error? Report it.