Glossary · Term

enumeration

← all terms

Definition

Plain language

Listing out every item that belongs in a group, one by one.

As stated in the literature

Explicitly emitting all members of a set, as opposed to characterizing it with a rule; for autoregressive models this couples member selection with an uncalibrated stopping decision.

Also called: enumerate, enumerating, enumerated

Why it matters: Listing forces two separate judgments — what to include and when to stop — and the stopping judgment is the one models are worst at, so lists come back short.

For example, instead of saying "all the even numbers under ten," enumeration means writing out 2, 4, 6, 8 and then deciding, on your own, that you are done.

Heard on the show

“Then they enumerate the top hundred and bottom hundred configurations by predicted score, run them for real, and keep the best.”
Episode 243 — How a Hundred Meaningless Word Choices Add Up to Flip a Model's Answer

Mentioned in 11 episodes

  1. 243
    How a Hundred Meaningless Word Choices Add Up to Flip a Model's Answer
  2. 233
    Why a Model Can Grade an Answer But Not Write the Answer Key
  3. 162
    The Empty-Lake Proof: Why More Rollouts Stop Helping Reasoning Models
  4. 157
    When an AI Coding Agent Drives a Phone Through the Terminal, No Screen Needed
  5. 146
    How an Innocent README Can Freeze an AI Agent's Safety Check for an Hour
  6. 132
    The Agent Failed — But Did the Instructions Deserve to Be Followed?
  7. 079
    An Old Idea From Cognitive Psychology Reshapes How We Reward Reasoning Models
  8. 062
    Treating Hallucinations as Exploits: A Gate-Based Architecture for Agent Safety
  9. 052
    An Old Reinforcement Learning Tradeoff Sneaks Back Into LLM Agents
  10. 042
    An Agentic Scientific Computing System That Actually Remembers What It Learns
  11. 034
    Catching Multi-Agent Deadlocks Before Deployment With a 40-Year-Old Tool

Related terms