# How Can We Synthesize High-Quality Pretraining Data? A Systematic Study of Prompt Design, Generator Model, and Source Data

arXiv: https://arxiv.org/abs/2604.13977 · HTML: https://mlsanity.com/paper/2604.13977/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- Builds on 39 of its 74 references that are in the map.
- First posted on arXiv 15 Apr 2026.

## Builds on (most cited first, 30 of 39)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://mlsanity.com/paper/1910.10683/) (2019)
- [Training Verifiers to Solve Math Word Problems](https://mlsanity.com/paper/2110.14168/) (2021)
- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [Scaling Laws for Neural Language Models](https://mlsanity.com/paper/2001.08361/) (2020)
- [Efficient Memory Management for Large Language Model Serving with PagedAttention](https://mlsanity.com/paper/2309.06180/) (2023)
- [Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge](https://mlsanity.com/paper/1803.05457/) (2018)
- [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://mlsanity.com/paper/2104.09864/) (2021)
- [HellaSwag: Can a Machine Really Finish Your Sentence?](https://mlsanity.com/paper/1905.07830/) (2019)
- [TriviaQA: A Large Scale Distantly Supervised Challenge Dataset for Reading Comprehension](https://mlsanity.com/paper/1705.03551/) (2017)
- [PIQA: Reasoning about Physical Commonsense in Natural Language](https://mlsanity.com/paper/1911.11641/) (2019)
- [FlashAttention-2: Faster Attention with Better Parallelism and Work Partitioning](https://mlsanity.com/paper/2307.08691/) (2023)
- [The Pile: An 800GB Dataset of Diverse Text for Language Modeling](https://mlsanity.com/paper/2101.00027/) (2020)
- [Can a Suit of Armor Conduct Electricity? A New Dataset for Open Book Question Answering](https://mlsanity.com/paper/1809.02789/) (2018)
- [Qwen2 Technical Report](https://mlsanity.com/paper/2407.10671/) (2024)
- [Know What You Don't Know: Unanswerable Questions for SQuAD](https://mlsanity.com/paper/1806.03822/) (2018)
- [GQA: Training Generalized Multi-Query Transformer Models from Multi-Head Checkpoints](https://mlsanity.com/paper/2305.13245/) (2023)
- [The FineWeb Datasets: Decanting the Web for the Finest Text Data at Scale](https://mlsanity.com/paper/2406.17557/) (2024)
- [Aligning Books and Movies: Towards Story-like Visual Explanations by Watching Movies and Reading Books](https://mlsanity.com/paper/1506.06724/) (2015)
- [DROP: A Reading Comprehension Benchmark Requiring Discrete Reasoning Over Paragraphs](https://mlsanity.com/paper/1903.00161/) (2019)
- [Phi-4 Technical Report](https://mlsanity.com/paper/2412.08905/) (2024)
- [Textbooks Are All You Need](https://mlsanity.com/paper/2306.11644/) (2023)
- [Compositional Semantic Parsing on Semi-Structured Tables](https://mlsanity.com/paper/1508.00305/) (2015)
- [Scaling Data-Constrained Language Models](https://mlsanity.com/paper/2305.16264/) (2023)
- [The Vendi Score: A Diversity Evaluation Metric for Machine Learning](https://mlsanity.com/paper/2210.02410/) (2022)
- [Are We Done with MMLU?](https://mlsanity.com/paper/2406.04127/) (2024)
- [Rephrasing the Web: A Recipe for Compute and Data-Efficient Language Modeling](https://mlsanity.com/paper/2401.16380/) (2024)
- [Nemotron-CC: Transforming Common Crawl into a Refined Long-Horizon Pretraining Dataset](https://mlsanity.com/paper/2412.02595/) (2024)
- [Common Sense Beyond English: Evaluating and Improving Multilingual Language Models for Commonsense Reasoning](https://mlsanity.com/paper/2106.06937/) (2021)
- [Instruction Pre-Training: Language Models are Supervised Multitask Learners](https://mlsanity.com/paper/2406.14491/) (2024)

## Built on by (most cited first, 6 of 6)

- [Deep Research Pretraining via Predictive Navigation](https://mlsanity.com/paper/2608.00432/) (2026)
- [DataOrchestra: Learning to Orchestrate Per-Example Curation of Pretraining Data](https://mlsanity.com/paper/2607.24717/) (2026)
- [The Announcement Carries the Cue: Markup, Boundaries, and the Notation of Pre-Training Corpora](https://mlsanity.com/paper/2608.09093/) (2026)
- [Where Does the Signal Live? A Web Data Recipe for Medical Encoder Pretraining](https://mlsanity.com/paper/2606.22079/) (2026)
- [Synthetic Pre-pretraining Survives Scale, but Not as a Grammatical Prior](https://mlsanity.com/paper/2609.39827/) (2026)
- [Effective Synthetic Data Curation Requires Group-Level Signals](https://mlsanity.com/paper/2610.00779/) (2026)

## Abstract (arXiv)

Synthetic data is a standard component in training large language models, yet systematic comparisons across design dimensions, including rephrasing strategy, generator model, and source data, remain absent. We conduct extensive controlled experiments, generating over one trillion tokens, to identify critical factors in rephrasing web text into synthetic pretraining data. Our results reveal that structured output formats, such as tables, math problems, FAQs, and tutorials, consistently outperform both curated web baselines and prior synthetic methods. Notably, increasing the size of the generator model beyond 1B parameters provides no additional benefit. Our analysis also demonstrates that the selection of the original data used for mixing substantially influences performance. By applying our findings, we develop textbftextscFinePhrase, a 486-billion-token open dataset of rephrased web text. We show that textscFinePhrase outperforms all existing synthetic data baselines while reducing generation costs by up to 30 times. We provide the dataset, all prompts, and the generation framework to the research community.