# Synthetic Rewriting as a Quality Multiplier: Evidence from Portuguese Continued Pretraining

arXiv: https://arxiv.org/abs/2603.24826 · HTML: https://mlsanity.com/paper/2603.24826/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 17 of its 19 references that are in the map.
- First posted on arXiv 25 Mar 2026.

## Builds on (most cited first, 17 of 17)

- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [TinyLlama: An Open-Source Small Language Model](https://mlsanity.com/paper/2401.02385/) (2024)
- [BLEnD: A Benchmark for LLMs on Everyday Knowledge in Diverse Cultures and Languages](https://mlsanity.com/paper/2406.09948/) (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)
- [INCLUDE: Evaluating Multilingual Language Understanding with Regional Knowledge](https://mlsanity.com/paper/2411.19799/) (2024)
- [All Languages Matter: Evaluating LMMs on Culturally Diverse 100 Languages](https://mlsanity.com/paper/2411.16508/) (2024)
- [Recycling the Web: A Method to Enhance Pre-training Data Quality and Quantity for Language Models](https://mlsanity.com/paper/2506.04689/) (2025)
- [Cabrita: closing the gap for foreign languages](https://mlsanity.com/paper/2308.11878/) (2023)
- [Rewriting Pre-Training Data Boosts LLM Performance in Math and Code](https://mlsanity.com/paper/2505.02881/) (2025)
- [Sabi'a: Portuguese Large Language Models](https://mlsanity.com/paper/2304.07880/) (2023)
- [ptt5-v2: A Closer Look at Continued Pretraining of T5 Models for the Portuguese Language](https://mlsanity.com/paper/2406.10806/) (2024)
- [TiEBe: Tracking Language Model Recall of Notable Worldwide Events Through Time](https://mlsanity.com/paper/2501.07482/) (2025)
- [Building High-Quality Datasets for Portuguese LLMs: From Common Crawl Snapshots to Industrial-Grade Corpora](https://mlsanity.com/paper/2509.08824/) (2025)
- [PoETa v2: Toward More Robust Evaluation of Large Language Models in Portuguese](https://mlsanity.com/paper/2511.17808/) (2025)
- [Curi'o-Edu 7B: Examining Data Selection Impacts in LLM Continued Pretraining](https://mlsanity.com/paper/2512.12770/) (2025)
- [Tucano 2 Cool: Better Open Source LLMs for Portuguese](https://mlsanity.com/paper/2603.03543/) (2026)

## Abstract (arXiv)

Synthetic data generation through document rewriting has emerged as a promising technique for improving language model pretraining, yet most studies focus on English and do not systematically control for the quality of the source data being rewritten. We present a controlled study of how synthetic rewriting interacts with source data quality in the context of Portuguese continued pretraining. Starting from ClassiCC-PT, a Portuguese corpus annotated with STEM and Educational quality scores, we construct two 10B-token subsets at different quality levels and rewrite each into four styles using a 7B instruction-tuned model, producing approximately 40B tokens of synthetic data per condition. We train two English-centric base models (1.1B and 7B parameters) on each condition and evaluate on PoETa V2, a comprehensive 44-task Portuguese benchmark. At the 7B scale, rewriting high-quality data yields a +3.4 NPM gain over the same data unmodified, while rewriting low-quality data provides only +0.5 NPM. At the 1.1B scale, this interaction is weaker, with unmodified low-quality data performing comparably to rewritten high-quality data. Our results demonstrate that synthetic rewriting acts primarily as a quality multiplier rather than a substitute for data curation, and that this effect is scale-dependent.