Non-determinism is the property of large language models that causes the same prompt to produce different responses when run multiple times. This is not a bug — it is a fundamental characteristic of how LLMs generate text by sampling from probability distributions at each token.Documentation Index
Fetch the complete documentation index at: https://docs.getcited.in/llms.txt
Use this file to discover all available pages before exploring further.
Why it matters
Non-determinism makes single-run AI visibility measurements unreliable. A brand mentioned in one ChatGPT response might not appear in the next response to the identical prompt. Robust measurement requires multiple runs per query and statistical aggregation — a single snapshot tells you what happened once, not what typically happens.How Cited uses it
Cited accounts for non-determinism by running prompts multiple times per platform and aggregating before reporting. Dashboard metrics are based on aggregated data, not single runs. The Cited Index achieves statistical stability through large sample sizes (253 brands across 185 queries). See Non-determinism for practical guidance on interpreting metric movements.Related concepts
- Non-determinism (full page) — detailed explanation and practical guidance
- Data freshness — how to interpret confidence in aggregated data
- LLM — the models exhibiting this property