README.md

metadata

license: cc-by-nc-sa-4.0
size_categories:
  - 10K<n<100K
task_categories:
  - question-answering
language:
  - en
tags:
  - arxiv
pretty_name: Arxiver Dataset with Category

Arxiver Dataset with Category

This dataset is an enhanced version of the Arxiver dataset with additional category information. It contains 63,357 arXiv papers published between January 2023 and October 2023, with each paper enriched with its primary category and all associated categories.

Additional Features

Beyond the original dataset's fields (ID, title, abstract, authors, publication date, URL, and markdown content), this version adds:

• primary_category: The main arXiv category of the paper (e.g., 'cs.AI', 'cs.CL')
• categories: A list of all arXiv categories associated with the paper

These category fields were obtained through the arXiv API to provide more comprehensive paper classification information.

Dataset Structure

{
    'id': 'arxiv paper id',
    'title': 'paper title',
    'abstract': 'paper abstract',
    'authors': 'paper authors',
    'published_date': 'publication date',
    'link': 'arxiv url',
    'markdown': 'paper content in markdown format',
    'primary_category': 'main arxiv category',
    'categories': ['list', 'of', 'arxiv', 'categories']
}

Using Arxiver

You can easily download and use the arxiver dataset with HF中国镜像站's datasets library.

from datasets import load_dataset
dataset = load_dataset("real-jiakai/arxiver-with-category")

Alternatively, you can stream the dataset to save disk space or to partially download the dataset:

from datasets import load_dataset
dataset = load_dataset("real-jiakai/arxiver-with-category", streaming=True)

License & Citation

This dataset follows the same Creative Commons Attribution-Noncommercial-ShareAlike (CC BY-NC-SA 4.0) license as the original Arxiver dataset.

References

When using this dataset, please cite both this version and the original Arxiver dataset:

@misc{acar_arxiver2024,
  author = {Alican Acar, Alara Dirik, Muhammet Hatipoglu},
  title = {ArXiver},
  year = {2024},
  publisher = {HF中国镜像站},
  howpublished = {\url{https://huggingface.co/datasets/neuralwork/arxiver}}
}

@misc{jiakai_arxiver2025,
  author = {real-jiakai},
  title = {Arxiver with Category},
  year = {2025},
  publisher = {HF中国镜像站},
  howpublished = {\url{https://huggingface.co/datasets/real-jiakai/arxiver-with-category}}
}