Mastery: connect the pieces

A complete explanation of Temperature and creativity has four parts. First: Temperature changes how sharply a model samples from its next-token probability distribution. Lower values favor high-probability continuations; higher values flatten differences and increase variation. Temperature does not add knowledge, remove bias, or reliably control factuality. Second, the mechanism: For logits z and temperature T, sampling uses softmax(z/T). When T is below 1, logit differences become larger and the distribution concentrates. Above 1, differences shrink and less-likely tokens receive more probability. At T approaching zero, implementations approximate greedy selection. Seeds and provider infrastructure may still prevent exact reproducibility. Third, the operational controls: temperature; top-p; top-k where available; random seed; number of candidates; prompt constraints; maximum tokens; and a separate selection or ranking step. Fourth, the evidence: For deterministic tasks, measure exact match, schema validity, variance, and repeatability over many runs. For ideation, measure distinct useful ideas, human preference, novelty without duplication, safety, and selection cost.

Use the scenario as an oral exam: For extracting a country code, use a strict schema and low temperature. For naming a gardening app, sample 20 candidates at a moderate temperature, deduplicate them, then score memorability and trademark risk. Defend one design choice, then argue against it using this tradeoff: Low temperature improves consistency but can repeat bland or systematically wrong answers. High temperature produces diverse candidates but raises variance and review cost. Combining high temperature with broad top-p can make outputs erratic; tune one sampling control at a time. Finally, identify which of these failures your design catches and which remain: calling temperature a creativity percentage, expecting temperature zero to guarantee identical responses, using high temperature for exact extraction, lowering temperature to fix missing evidence, changing temperature and top-p simultaneously, and evaluating one sample instead of a distribution.

Mastery is not recalling every term. It is predicting consequences before running the system, noticing when evidence contradicts the prediction, and revising the design without moving the goalposts. Keep a decision record containing the workload, baseline, configuration, test set version, results, known limitations, owner, and rollback condition.

For extracting a country code, use a strict schema and low temperature. For naming a gardening app, sample 20 candidates at a moderate temperature, deduplicate them, then score memorability and trademark risk.

The worked number is with logits [2,1], T=1 gives probabilities about [0.73,0.27]; T=0.5 gives [0.88,0.12], showing concentration rather than a linear creativity dial. State the unit and denominator whenever you report it. A percentage without a denominator can conceal a tiny sample; a latency without a percentile can conceal slow users; a similarity score without a labeled task can conceal irrelevant neighbors. Compare the observed value with a threshold chosen before seeing the final test result.

Now test the tempting shortcut. Suppose the team optimizes only the most visible metric. The result may look better while the system becomes less trustworthy. The reason is concrete: Low temperature improves consistency but can repeat bland or systematically wrong answers. High temperature produces diverse candidates but raises variance and review cost. Combining high temperature with broad top-p can make outputs erratic; tune one sampling control at a time. This is why the decision record must include both the intended gain and the tolerated regression. If the tolerated regression is unknown, the change is not ready for a consequential workflow.

• Prefer a measured baseline over a persuasive demo.
• Keep versions, inputs, and thresholds reproducible.
• Separate syntactic success from semantic correctness and authorization.
• Escalate or abstain when evidence falls outside the contract.
• Re-evaluate when data, traffic, models, providers, or user goals change.

These rules turn the topic into an engineering decision rather than a slogan. They also make disagreement productive: another person can challenge the assumptions, rerun the evaluation, and reach a documented conclusion.

Give a three-minute teach-back with no slides. Minute one: define the technique and its boundary. Minute two: trace the mechanism using the worked case and calculation. Minute three: defend the chosen controls with evaluation evidence, then name the strongest unresolved failure. Ask the listener to change one assumption and update your recommendation aloud. You have mastered the topic when the recommendation changes for a technical reason—not because the vocabulary changed—and when you can specify the next experiment that would reduce the most consequential uncertainty.