Characterizing Model Behavior Under Synthetic Data Training: An Empirical Study Across Scales and Mixing Ratios

• URL: http://arxiv.org/abs/2510.05133v1
• Date: Wed, 01 Oct 2025 03:28:01 GMT
• Title: Characterizing Model Behavior Under Synthetic Data Training: An Empirical Study Across Scales and Mixing Ratios
• Authors: Y. Du, G. Wu, G. Tang, W. Wang, Q. Fan,
• Abstract summary: This paper presents a controlled empirical study examining model performance, calibration, and output characteristics when trained on varying synthetic-to-external data ratios.<n>Our key findings include: models maintain stable performance with up to 20% synthetic data, but degradation accelerates beyond 30%.<n>Current best practices, such as those employed in STaR and Self-Instruct systems that maintain greater than 80% external data, operate well within safe regimes identified by our experiments.
• Score: 1.631115063641726
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Synthetic data generated by large language models has become integral to modern NLP training pipelines, from bootstrapping reasoning capabilities to augmenting instruction-following datasets. While recent work demonstrates successful applications maintaining high external data ratios, systematic understanding of how synthetic data proportion affects model behavior across different scales remains limited. This paper presents a controlled empirical study examining model performance, calibration, and output characteristics when trained on varying synthetic-to-external data ratios. Using the Pythia model suite (410M-12B parameters) across five diverse tasks, we evaluate models after one to three training iterations with synthetic data proportions ranging from 0-50\%. Our key findings include: models maintain stable performance with up to 20\% synthetic data, but degradation accelerates beyond 30\%; larger models (6.9B-12B) show greater robustness to synthetic data than smaller models (410M-1.4B); calibration degradation precedes accuracy loss, providing an early warning signal; and task characteristics matter, with reasoning tasks degrading faster than retrieval tasks under synthetic data training. Importantly, we find that current best practices, such as those employed in STaR and Self-Instruct systems that maintain greater than 80\% external data, operate well within safe regimes identified by our experiments. We provide practical guidance for practitioners on synthetic data budgets based on model scale and task requirements, alongside detailed comparison with concurrent work including Shumailov et al.'s model collapse findings.

Related papers

• Demystifying Synthetic Data in LLM Pre-training: A Systematic Study of Scaling Laws, Benefits, and Pitfalls [25.294408301653576]
  Training data plays a crucial role in Large Language Models (LLM) scaling, yet high quality data is of limited supply.<n>We compare natural web data, diverse synthetic types (rephrased text, generated textbooks), and mixtures of natural and synthetic data.<n>We find pre-training on rephrased synthetic data textitalone is not faster than pre-training on natural web texts.
  arXiv  Detail & Related papers  (2025-10-02T03:24:42Z)
• DAViD: Data-efficient and Accurate Vision Models from Synthetic Data [6.829390872619486]
  We demonstrate that it is possible to train models on much smaller but high-fidelity synthetic datasets.<n>Our models require only a fraction of the cost of training and inference when compared with foundational models of similar accuracy.
  arXiv  Detail & Related papers  (2025-07-21T08:17:41Z)
• Scaling Laws of Synthetic Data for Language Models [125.41600201811417]
  We introduce SynthLLM, a scalable framework that transforms pre-training corpora into diverse, high-quality synthetic datasets.<n>Our approach achieves this by automatically extracting and recombining high-level concepts across multiple documents using a graph algorithm.
  arXiv  Detail & Related papers  (2025-03-25T11:07:12Z)
• A Theoretical Perspective: How to Prevent Model Collapse in Self-consuming Training Loops [55.07063067759609]
  High-quality data is essential for training large generative models, yet the vast reservoir of real data available online has become nearly depleted.<n>Models increasingly generate their own data for further training, forming Self-consuming Training Loops (STLs)<n>Some models degrade or even collapse, while others successfully avoid these failures, leaving a significant gap in theoretical understanding.
  arXiv  Detail & Related papers  (2025-02-26T06:18:13Z)
• Golden Ratio Weighting Prevents Model Collapse [7.512957145774808]
  We develop an optimal training strategy for integrating real and synthetic data.<n>We characterize the impact of the mixing proportion and weighting scheme of synthetic data on the final model's performance.<n>In some cases, the optimal weight assigned to real data corresponds to the reciprocal of the golden ratio.
  arXiv  Detail & Related papers  (2025-02-25T10:15:16Z)
• Little Giants: Synthesizing High-Quality Embedding Data at Scale [71.352883755806]
  We introduce SPEED, a framework that aligns open-source small models to efficiently generate large-scale embedding data. SPEED uses only less than 1/10 of the GPT API calls, outperforming the state-of-the-art embedding model E5_mistral when both are trained solely on their synthetic data.
  arXiv  Detail & Related papers  (2024-10-24T10:47:30Z)
• Collapse or Thrive? Perils and Promises of Synthetic Data in a Self-Generating World [19.266191284270793]
  generative machine learning models are pretrained on web-scale datasets containing data generated by earlier models.<n>Some prior work warns of "model collapse" as the web is overwhelmed by synthetic data.<n>We report experiments on three ways of using data (training-workflows) across three generative model task-settings.
  arXiv  Detail & Related papers  (2024-10-22T05:49:24Z)
• Unveiling the Flaws: Exploring Imperfections in Synthetic Data and Mitigation Strategies for Large Language Models [89.88010750772413]
  Synthetic data has been proposed as a solution to address the issue of high-quality data scarcity in the training of large language models (LLMs) Our work delves into these specific flaws associated with question-answer (Q-A) pairs, a prevalent type of synthetic data, and presents a method based on unlearning techniques to mitigate these flaws. Our work has yielded key insights into the effective use of synthetic data, aiming to promote more robust and efficient LLM training.
  arXiv  Detail & Related papers  (2024-06-18T08:38:59Z)
• Let's Synthesize Step by Step: Iterative Dataset Synthesis with Large Language Models by Extrapolating Errors from Small Models [69.76066070227452]
  *Data Synthesis* is a promising way to train a small model with very little labeled data. We propose *Synthesis Step by Step* (**S3**), a data synthesis framework that shrinks this distribution gap. Our approach improves the performance of a small model by reducing the gap between the synthetic dataset and the real data.
  arXiv  Detail & Related papers  (2023-10-20T17:14:25Z)
• Does Synthetic Data Make Large Language Models More Efficient? [0.0]
  This paper explores the nuances of synthetic data generation in NLP. We highlight its advantages, including data augmentation potential and the introduction of structured variety. We demonstrate the impact of template-based synthetic data on the performance of modern transformer models.
  arXiv  Detail & Related papers  (2023-10-11T19:16:09Z)
• SynBench: Task-Agnostic Benchmarking of Pretrained Representations using Synthetic Data [78.21197488065177]
  Recent success in fine-tuning large models, that are pretrained on broad data at scale, on downstream tasks has led to a significant paradigm shift in deep learning. This paper proposes a new task-agnostic framework, textitSynBench, to measure the quality of pretrained representations using synthetic data.
  arXiv  Detail & Related papers  (2022-10-06T15:25:00Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.