Quantifying non deterministic drift in large language models
Abstract
Large language models (LLMs) are widely used for tasks ranging from summarisation to decision support. In practice, identical prompts do not always produce identical outputs, even when temperature and other decoding parameters are fixed. In this work, we conduct repeated-run experiments to empirically quantify baseline behavioural drift, defined as output variability observed when the same prompt is issued multiple times under operator-free conditions. We evaluate two publicly accessible models, gpt-4o-mini and llama3.1-8b, across five prompt categories using exact repeats, perturbed inputs, and reuse modes at temperatures of 0.0 and 0.7. Drift is measured using unique output fractions, lexical similarity, and word count statistics, enabling direct comparison across models, prompting modes, and deployment types. The results show that nondeterminism persists even at temperature 0.0, with distinct variability patterns by model size, deployment, and prompt type. We situate these findings within existing work on concept drift, behavioural drift, and infrastructure-induced nondeterminism, discuss the limitations of lexical metrics, and highlight emerging semantic approaches. By establishing a systematic empirical baseline in the absence of stabilisation techniques, this study provides a reference point for evaluating future drift mitigation and control methods.
Cite
@article{arxiv.2601.19934,
title = {Quantifying non deterministic drift in large language models},
author = {Claire Nicholson},
journal= {arXiv preprint arXiv:2601.19934},
year = {2026}
}
Comments
10 pages, 3 figures, 1 table. Empirical measurement study reporting new repeated-run experiments quantifying baseline nondeterministic drift in large language models. This manuscript presents original empirical results (not a review or position paper) and establishes a baseline reference for future drift-mitigation work