Datasets:
Tasks:
Text Generation
Modalities:
Text
Formats:
parquet
Languages:
English
Size:
10K - 100K
License:
| license: cc-by-4.0 | |
| task_categories: | |
| - text-generation | |
| language: | |
| - en | |
| tags: | |
| - cosmology | |
| - arxiv | |
| - latex | |
| - physics | |
| - astronomy | |
| size_categories: | |
| - 10K<n<100K | |
| # Cosmology arXiv Papers (astro-ph.CO) | |
| LaTeX source files for cosmology papers from arXiv. | |
| ## Dataset Description | |
| - **Papers**: 73,487 cosmology papers | |
| - **Date range**: 2007-2025 | |
| - **Category**: astro-ph.CO | |
| ## Columns | |
| | Column | Description | | |
| |--------|-------------| | |
| | `arxiv_id` | arXiv paper ID | | |
| | `title` | Extracted paper title | | |
| | `abstract` | Extracted abstract | | |
| | `year` | Publication year | | |
| | `tex_source` | Full LaTeX source (with \input expanded) | | |
| ## Usage | |
| ```python | |
| from datasets import load_dataset | |
| ds = load_dataset("adammoss/cosmology-arxiv") | |
| # Access a paper | |
| paper = ds['train'][0] | |
| print(paper['title']) | |
| print(paper['abstract']) | |
| ``` | |
| ## Preprocessing | |
| The `tex_source` contains raw LaTeX with `\input{}` and `\include{}` commands expanded. | |
| You may want to apply additional preprocessing depending on your use case. | |
| ## License | |
| The dataset follows arXiv's license terms. Most papers are CC-BY or similar. | |
| See individual papers for specific licensing. | |
| ## Updates | |
| This dataset is updated periodically with new papers. | |
| Last updated: 2025-12-12 | |