Temperature

Temperature is a sampling parameter that controls the randomness of a language model's output. It scales the token probability distribution before sampling: values near zero make output nearly deterministic and focused, while higher values flatten the distribution, producing more diverse and creative but less predictable text.

Definition

Temperature

At each generation step a model produces a probability for every possible next token. Temperature divides the underlying logits before they are turned into probabilities. A low temperature sharpens the distribution so the most likely tokens dominate, favoring consistency and factual tasks; a high temperature flattens it, giving lower-probability tokens a real chance and increasing variety.

Temperature is often tuned alongside top-p (nucleus) sampling. Providers generally recommend adjusting one or the other rather than both at once, since they interact. For tasks that need reproducibility, a temperature at or near zero is typical.

Frequently asked questions

What is Temperature?

Temperature is a sampling parameter that controls the randomness of a language model's output. It scales the token probability distribution before sampling: values near zero make output nearly deterministic and focused, while higher values flatten the distribution, producing more diverse and creative but less predictable text.

Sources

Related terms

← Full AI & prompt engineering glossary

Put Temperature to work with Prompeteer, the Agentic Contextual AI Platform →