- A technique in which a large language model (LLM) is provided with a small number of examples (typically 2–10) of a task before being asked to perform similar tasks on new inputs. This method leverages the model’s ability to recognize patterns from limited examples to generalize to new instances.
This term is sponsored by: your name/company?
- Previous term: Fetch priority
- Next term: FFmpeg
- Random term: Rubbish in, rubbish out (webglossary.info/random 🎲)