Glossary · Term

temperature

← all terms

Definition

Plain language

A knob that controls how creative or how predictable a model's output is.

As stated in the literature

A scalar that scales logits before softmax during sampling; higher values flatten the distribution and produce more diverse outputs, lower values make generation more deterministic.

Why it matters: It is the simplest knob for trading off creativity against reliability, and the right setting depends heavily on whether the task is open-ended or factual.

For example, setting temperature to 0 makes the model pick the single most likely next word every time, while a temperature of 1 lets it sample more freely.

Heard on the show

“A reagent, a quantity, a temperature, or the order of two steps.”
Episode 244 — The Open-Weight Defense That Feeds Attackers Confident, Falsified Answers

Mentioned in 31 episodes

  1. 244
    The Open-Weight Defense That Feeds Attackers Confident, Falsified Answers
  2. 240
    Frontier Models Designed Follow-Ups To Fraudulent Papers 93% Of The Time
  3. 238
    How a Cheap Model Reads the Flagship's Secret Reasoning Aloud
  4. 234
    Two Copies of Gemini Cooperated in a Game Where Betrayal Always Pays
  5. 230
    Why AI Survey Panels Break Before the Dice Ever Roll
  6. 228
    Same Chatbot, Two Doors: Why 'Grok's Opinion' Doesn't Exist
  7. 225
    How a Frozen Model Went From Zero to Sixty Percent by Borrowing Another's Thinking
  8. 219
    Forty-Four AI Models, One Word, And The Newest Ones Conform Most
  9. 216
    The AI Tutor That Gives Poor Kids a Thinner History
  10. 215
    The Same Policy Scored 85 for the US and 36 for Russia
  11. 197
    Twin Problems Suggest AI Reasoning Gains Are Mostly Better Fact Recall
  12. 195
    Why 'Be Careful' Does Nothing for AI Coding Agents, and What Does
  13. 179
    How DeepSeek Made One User Faster Without Slowing Down the Crowd
  14. 174
    When the AI 'Schemes,' It's Usually Just Lazy or Confused
  15. 162
    The Empty-Lake Proof: Why More Rollouts Stop Helping Reasoning Models
  16. 149
    When Cornering a Chatbot Makes It Lie: J.P. Morgan's Case for 'Playing Dead'
  17. 148
    Why Letting an AI Watch Its Own Scoreboard Can Quietly Overwrite Its Safety
  18. 133
    How MiniMax Turned a Reward-Hacking Disaster Into Olympiad Gold
  19. 125
    AI Coding Agents Run a Marathon, and Fewer Than One in Three Finish
  20. 106
    Giving Agents a Notebook Instead of New Weights: How ExpGraph Lets Frozen Models Learn
  21. 098
    Finding Millions of Readable Concepts Inside a Real, Deployed AI Model
  22. 093
    A Calibrated Knob for Weak-to-Strong AI Oversight, Tested on Real Code
  23. 073
    When Three LLMs Talk to Each Other, Their Ideas Quietly Stop Moving
  24. 072
    A Robot Made Graphene Without Help, And Caught Itself Hallucinating
  25. 070
    When Models Know the Answer But Say the Wrong Thing Anyway
  26. 055
    Why LLM Judges Flip Their Verdicts When You Change the Question Format
  27. 023
    Why a Small Agent Confidently Overwrites Memories It Doesn't Understand
  28. 018
    Language Models Compute the Rational Move, Then Override It
  29. 015
    The Audit Number Isn't What You Think: Sycophancy and the Case Against Single-Prompt Bias Tests
  30. 013
    Why Search Keeps Rediscovering the Same Workflow, and What That Means
  31. 011
    When RL Actually Teaches Agents Something New, And When It Doesn't

Related terms

logits softmax