Skip to content

Overview

Given a prompt, the model will return one or more predicted completions, and can also return the probabilities of alternative tokens at each position.