Rewriting Pre-Training Data Boosts LLM Performance in Math and Code

TL;DR

This paper demonstrates that systematically rewriting pre-training data with a novel pipeline significantly improves large language models' performance in math and code tasks, surpassing previous baselines.

Contribution

Introduces two openly licensed datasets, SwallowCode and SwallowMath, with a transform-and-retain pipeline that enhances data quality for LLM pre-training, leading to substantial performance gains.

Findings

SwallowCode improves code generation metrics by over 16 points.

SwallowMath increases accuracy on GSM8K and MATH benchmarks.

Pipeline ablations show each stage contributes to overall performance gains.

Abstract

The performance of large language models (LLMs) in program synthesis and mathematical reasoning is fundamentally limited by the quality of their pre-training corpora. We introduce two openly licensed pre-training datasets, released under the Llama 3.3 Community License, that significantly enhance LLM performance by systematically rewriting public data. SwallowCode ($\approx$16.1 billion tokens) refines Python snippets from The-Stack-v2 through a novel four-stage pipeline: syntax validation, pylint-based style filtering, and a two-stage LLM rewriting process that enforces style conformity and transforms snippets into self-contained, algorithmically efficient examples. Unlike prior methods that rely on exclusionary filtering or limited transformations, our transform-and-retain approach refines low-quality code, maximizing data utility. SwallowMath ($\approx$2.3 billion tokens) enhances Finemath-4+ by removing boilerplate, restoring context, and reformatting solutions into concise, step-by-step explanations. Within a fixed 50 billion token training budget, continual pre-training of Llama-3.1-8B with SwallowCode boosts pass@1 by +17.0 on HumanEval and +16.1 on HumanEval+ compared to Stack-Edu, surpassing the baseline model's code generation capabilities. Similarly, substituting SwallowMath yields +12.4 accuracy on GSM8K and +7.6 on MATH. Ablation studies confirm that each pipeline stage contributes incrementally, with rewriting yielding the largest gains. By releasing datasets, prompts, checkpoints, and pipeline code, we ensure reproducibility and provide a transferable transform-and-retain methodology that can be adapted to other base models and LLM rewriting setups.