Glossary
Temperature
A sampling setting controlling output randomness.
What it is
Temperature scales how randomly a model samples the next token. Lower values give deterministic, focused output (good for extraction); higher values increase diversity (useful for brainstorming).
Why it matters in production
Temperature is a production decision per use case, not a default you inherit: near-zero for extraction, classification, and tool calls where consistency matters; higher only where you actually want variety.
In practice
Setting temperature to 0 for a JSON extraction pipeline so the same input yields the same parse, and 0.7 for a name-brainstorming helper.
Related terms
Production AI Notes
One practical AI engineering email each week
One concept, one architecture, one project idea, and one interview question — written for developers who want to build and ship real AI systems.