lucius1022/DeMix_CorporaDataset Card for DeMix Corpora DeMix 📄 Paper: Decouple Searching from Training: Scaling Data Mixing via Model Merging for Large Language Model Pre training 🤗 Dataset: DeMix Corpora 🐱 Github: Demix Dataset Details Dataset Description DeMix Corpora (15T original tokens and 22T mixture tokens) serves as a comprehensive, high quality, large scale, and carefully mixed resource that can be directly employed for pre training.… See the full description on the dataset page: https://huggingface.co/datasets/lucius1022/DeMix Corpora.

DeMix_Corpora

PUBLIC

Mirrored

Dataset Card for DeMix Corpora DeMix 📄 Paper: Decouple Searching from Training: Scaling Data Mixing via Model Merging for Large Language Model Pre-training 🤗 Dataset: DeMix Corpora 🐱 Github: Demix Dataset Details Dataset Description DeMix Corpora (15T original tokens and 22T mixture tokens) serves as a comprehensive, high-quality, large-scale, and carefully mixed resource that can be directly employed for pre-training.… See the full description on the dataset page: https://huggingface.co/datasets/lucius1022/DeMix_Corpora.

Formats:json, jsonlLibraries:Datasets, Dask, Polars, CroissantLicense:apache-2.0

ArXiv:2602.00747

imported

huggingface

minio

license:apache-2.0

size_categories:n<1K

modality:tabular

region:us

0downloads0likes16.8 MBLicense:apache-2.0

Dataset Card

Edit

No dataset card provided yet.

Formats

json

jsonl

License

apache-2.0

Source

Mirrored from an external registry.

Last synced 6/14/2026

Paper for this model

arXiv:2602.00747

Paper2602.00747 0

Dataset Viewer

API

Preview not yet available for this dataset