Few-Shot Prompting

Appears in 1 paper

A technique where a language model is shown a small number of examples (typically 2-8) before being asked to solve a new problem.

As used in Paper 14 — Chain-of-Thought Prompting Elicits Reasoning in Large Language Models →

A technique where a language model is shown a small number of examples (typically 2-8) before being asked to solve a new problem. The model learns the pattern from these examples and applies it to the test case. Unlike fine-tuning, few-shot prompting happens at inference time, with no model parameter updates. It's one form of in-context learning.