Lowering an LLM's temperature setting is one lever for improving output consistency¶

Claim · All claims
Evidence · 1 study · theoretical r2 · q2
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person · recheck by 2029: tool and model accuracy changes quickly

Subclaims¶

q2 i? Lowering the model's temperature setting, which controls output randomness, is one lever for improving consistency of AI outputs across repeated uses. → Li 2026

Evidence¶

Li 2026¶

Li, W., Kurimchak, M., & Whitmer, J. (2026). ATS Field Note: AI Model Evaluation. ATS Hub Field Note Series. https://doi.org/10.51388/20.500.12265/300

q2 · i? · theoretical · r2

In the reliability discussion, the field note identifies temperature as a lever for improving consistency, noting that "Some variability is expected with Large Language Models (LLMs)". No experiment or effect size is reported.

"Lowering the model's "temperature" setting, which controls output randomness, is one lever for improving consistency."

Discussion¶

Related Claims¶

• Same-model LLM configuration pairs agree more than cross-model pairs, and agreement decreases monotonically as temperature difference increases — related
• Prompt settings and temperature settings have significant main effects on correlations among the three LLM chatbots — related
• Effective AI use in nursing education depends on prompting skills, which participants identify as critical for obtaining accurate and reliable outputs — related
• Within-judge variance of the repeated LLM-judge rubric is negligible, but cross-model-family agreement remains untested — related
• Temperature-construct interactions: constructs with moderate theoretical coherence benefited from higher temperatures, while well-defined constructs required deterministic settings — a narrower finding that bears on this claim