On the Diversity of Synthetic Data and its Impact on Training Large Language Models

ICLR 2026 Conference Submission20393 Authors

Keywords: Data Diversity for LLM Pretraining, Synthetic Data

TL;DR: LLM Cluster-agent to measure synthetic data diversity and shows that greater diversity consistently improves LLM performance in both pre-training and fine-tuning.

Abstract:

The rise of Large Language Models (LLMs) has accentuated the need for diverse, high-quality pre-training data.
Synthetic data emerges as a viable solution to the challenges of data scarcity and inaccessibility.
While previous literature has focused predominantly on the quality and quantity of real data, our work enables the measurement of diversity in synthetic data and explores its impact on LLM performance.
We study the downstream effects of synthetic data diversity during both the pre-training and fine-tuning stages by introducing a new diversity metric, \textit{LLM cluster-agent}, designed to evaluate the diversity of synthetic datasets.
Through a series of controlled experiments with models of 350M and 1.4B parameters, we demonstrate that the proposed cluster-based LLM scoring of diversity correlates positively with both pre-training and supervised fine-tuning performance.
Our findings also reveal that synthetic data diversity in pre-training affects supervised fine-tuning more significantly than pre-training itself, even for smaller models.
We hope this study advances our understanding of the optimal use of synthetic data in LLM training and opens new avenues for efficient data generation processes.

Primary Area: foundation or frontier models, including LLMs

Submission Number: 20393