Temperature is the dial that controls how much randomness a language model introduces when picking its next token. At temperature zero, the model always picks the single most probable token - output is deterministic and maximally conservative. As temperature rises toward one and beyond, it draws from a wider range of less probable tokens, producing text that is more varied, surprising, and creative - but also less anchored to the evidence in front of it.
Why it matters
The name is borrowed from physics: in thermodynamics, higher temperature means more energy and more disorder. The analogy holds. A high-temperature model is more likely to drift from the source material, introduce detail that wasn't there, or follow an interesting tangent rather than the most faithful reading of the text.
For most document AI tasks - summarization, Q&A, extraction - that drift is a problem, not a feature. When the goal is a citable answer that faithfully reflects what a document actually says, you want the model focused and repeatable, not creative. That means running inference at a low temperature so the output tracks the source rather than embellishes it.
Sidenote generates answers at a low temperature by design, so output stays conservative and every claim can be pinned back to the exact passage it came from.
FAQ
What temperature should I use?
Low, roughly 0 to 0.3, for factual work: summaries, extraction, document Q&A, anywhere the output should track a source. Higher values, around 0.7 and up, suit brainstorming and creative writing, where variety matters more than fidelity.
Does temperature 0 guarantee the same answer every time?
Nearly, but not absolutely. At zero the model always picks the most probable token, which makes output as repeatable as inference gets, though hardware-level nondeterminism and provider-side model updates can still produce small variations between runs.