Skip to content


Data sources

RedPajama Pile CommonCrawl (webscrape) C4 (CommonCrawl) Github Books Arxiv StackExchange

The largest open source text dataset just dropped

Dolma. (by AI2)

WARNING: The license is not 'open source' 3 Trillion tokens of high quality data.

  • Diverse: Documents, code, academic papers, wiki..
  • Focused: English only.
  • De-duplicated.
  • Filtered for high quality.

But most importantly: The largest open curated dataset for pretraining.

ā€¢ Link: ā€¢ Blog: ā€¢ Code: ā€¢ Paper:

Process Supervision

MathPile: A Billion-Token-Scale Pretraining Corpus for Math

High-quality, large-scale corpora are the cornerstone of building foundation models. In this work, we introduce MathPile, a diverse and high-quality math-centric corpus comprising about 9.5 billion tokens. Throughout its creation, we adhered to the principle of ``less is more'', firmly believing in the supremacy of data quality over quantity, even in the pre-training phase. Our meticulous data collection and processing efforts included a complex suite of preprocessing, prefiltering, language identification, cleaning, filtering, and deduplication, ensuring the high quality of our corpus. Furthermore, we performed data contamination detection on downstream benchmark test sets to eliminate duplicates. We hope our MathPile can help to enhance the mathematical reasoning abilities of language models. We plan to open-source different versions of \mathpile with the scripts used for processing, to facilitate future developments in this field.

Multimodal Datasets

SPIQA: A Dataset for Multimodal Question Answering on Scientific Papers

Paper Data