Glossary · Term

structured output

← all terms

Definition

Plain language

Asking a model to answer in a fixed, machine-readable format instead of free-flowing text.

As stated in the literature

Constraining model generation to a specified schema (e.g. a JSON distribution over options), used here to elicit a described distribution rather than per-call samples.

Why it matters: This matters because a fixed, machine-readable format lets other programs reliably read and use the model's answers without guessing at messy free text.

For example, instead of a chatty paragraph, you might require the model to reply with a tidy list like {"yes": 60, "no": 40}.

Heard on the show

“And even describe-as-a-technique isn't new — a paper last year proposed asking the model for its distribution as structured output, and on this benchmark the two are statistically indistinguishable.”
Episode 230 — Why AI Survey Panels Break Before the Dice Ever Roll

Mentioned in 5 episodes

  1. 230
    Why AI Survey Panels Break Before the Dice Ever Roll
  2. 154
    How a 7B Model Out-Investigates a 72B One by Choosing What to Look At
  3. 100
    How a Prompt Wrapper Lets a Frontier Model Play Poker Like an Expert
  4. 082
    Training a Deep Research Agent on 8,000 Synthetic Tasks: The Rubric Tree Trick
  5. 023
    Why a Small Agent Confidently Overwrites Memories It Doesn't Understand

Related terms