Nemotron-CC: Transforming Common Crawl into a Refined Long-Horizon Pretraining Dataset

Result code in IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216208%3A11320%2F26%3ABKCY2XIR" target="_blank" >RIV/00216208:11320/26:BKCY2XIR - isvavai.cz</a>

Result on the web

<a href="https://aclanthology.org/2025.acl-long.123/" target="_blank" >https://aclanthology.org/2025.acl-long.123/</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.18653/v1/2025.acl-long.123" target="_blank" >10.18653/v1/2025.acl-long.123</a>

Result language

angličtina

Original language name

Nemotron-CC: Transforming Common Crawl into a Refined Long-Horizon Pretraining Dataset

Original language description

Recent English Common Crawl datasets like FineWeb-Edu and DCLM achieved significant benchmark gains via aggressive model-based filtering, but at the cost of removing 90% of data. This limits their suitability for long token horizon training, such as 15T tokens for Llama 3.1. In this paper, we show how to achieve better trade-offs between accuracy and data quantity by a combination of classifier ensembling, synthetic data rephrasing, and reduced reliance on heuristic filters. When training 8B parameter models for 1T tokens, using a high-quality subset of our data improves MMLU by 5.6 over DCLM, demonstrating the efficacy of our methods for boosting accuracies over a relatively short token horizon. Furthermore, our full 6.3T token dataset matches DCLM on MMLU, but contains four times more unique real tokens than DCLM. This unlocks state-of-the-art training over a long token horizon: an 8B parameter model trained for 15T tokens, of which 7.2T came from our dataset, is better than the Llama 3.1 8B model: +5 on MMLU, +3.1 on ARC-Challenge, and +0.5 on average across ten diverse tasks. The dataset is available at https://data.commoncrawl.org/contrib/Nemotron/Nemotron-CC/index.html.

Czech name

—

Czech description

—

Type

D - Article in proceedings

CEP classification

—

OECD FORD branch

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Project

—

Continuities

—

Publication year

2025

Confidentiality

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Article name in the collection

Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)

ISBN

979-8-89176-251-0

ISSN

—

e-ISSN

—

Number of pages

17

Pages from-to

2459-2475

Publisher name

Association for Computational Linguistics

Place of publication

—

Event location

Vienna, Austria

Event date

Jan 1, 2026

Type of event by nationality

WRD - Celosvětová akce

UT code for WoS article

—