Upload folder using huggingface_hub

.gitattributes

@@ -57,3 +57,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 # Video files - compressed
 *.mp4 filter=lfs diff=lfs merge=lfs -text
 *.webm filter=lfs diff=lfs merge=lfs -text
+data/arxiv_citations_valid.jsonl filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,137 @@
+---
+license: cc0-1.0
+task_categories:
+  - text-classification
+  - feature-extraction
+language:
+  - en
+tags:
+  - arxiv
+  - crossref
+  - citations
+  - scholarly-communications
+  - bibliometrics
+  - doi
+pretty_name: Crossref arXiv Citations
+size_categories:
+  - 100K<n<1M
+configs:
+  - config_name: default
+    data_files:
+      - split: train
+        path: data/arxiv_citations_valid.jsonl
+---
+
+# Crossref arXiv Citations
+
+A dataset of arXiv preprints and their citations extracted from Crossref metadata, validated against DataCite records and DOI resolution.
+
+## Dataset Description
+
+This dataset maps arXiv works to the scholarly works that cite them, extracted from the Crossref public data file. Each record represents an arXiv preprint with all known citations from Crossref-registered works.
+
+### Dataset Statistics
+
+| Metric | Value |
+|--------|-------|
+| arXiv works | 923,302 |
+| Total citations | 5,121,638 |
+| Max citations (single work) | 40,040 |
+
+## Data Schema
+
+Each record contains:
+
+| Field | Type | Description |
+|-------|------|-------------|
+| `arxiv_doi` | string | The DOI for the arXiv work (format: `10.48550/arXiv.{id}`) |
+| `arxiv_id` | string | The arXiv identifier (e.g., `1412.6980`) |
+| `citation_count` | integer | Number of citing works |
+| `cited_by` | array | List of citation objects |
+
+Each citation object in `cited_by` contains:
+
+| Field | Type | Description |
+|-------|------|-------------|
+| `doi` | string | DOI of the citing work |
+| `raw_match` | string | The original text that matched the arXiv reference |
+| `reference` | object | The full reference object from the citing work's metadata |
+
+### Example Record
+
+```json
+{
+  "arxiv_doi": "10.48550/arXiv.1412.6980",
+  "arxiv_id": "1412.6980",
+  "citation_count": 40040,
+  "cited_by": [
+    {
+      "doi": "10.1016/j.jhydrol.2025.134709",
+      "raw_match": "arXiv.1412.6980",
+      "reference": {
+        "key": "10.1016/j.jhydrol.2025.134709_b0175",
+        "unstructured": "Kingma, D.P., & Ba, J. (2014), Adam: A Method for Stochastic Optimization..."
+      }
+    }
+  ]
+}
+```
+
+## Extraction Process
+
+### 1. Reference Extraction
+
+arXiv references were extracted from the Crossref January 2025 Public Data File by scanning reference metadata for arXiv identifiers. The extraction recognizes multiple reference formats:
+
+- **arXiv IDs**: `arXiv:1412.6980`, `arXiv: 2206.15325`
+- **arXiv DOIs**: `10.48550/arXiv.1412.6980`
+- **arXiv URLs**: `arxiv.org/abs/1412.6980`
+- **Legacy format**: `arXiv:cs.DM/9910013`
+
+Identifiers are normalized (lowercase, version suffixes removed) and deduplicated per citing work.
+
+### 2. Validation
+
+The extracted arXiv DOIs were validated through a two-stage process:
+
+1. **DataCite matching**: DOIs were checked against the DataCite arXiv record set (~2.8M records)
+2. **DOI resolution**: For DOIs not found in DataCite, HTTP HEAD requests to `doi.org` verified resolution
+
+Only arXiv works with valid, resolvable DOIs are included in this dataset.
+
+## Data Sources
+
+- **Crossref Public Data File**: January 2025 snapshot
+- **DataCite arXiv Records**: Used for DOI validation
+
+## Use Cases
+
+- Citation analysis for arXiv preprints
+- Studying informal citation patterns (preprint vs. published version citations)
+- Bibliometric research on preprint impact
+- Training models for reference parsing and entity linking
+
+## Limitations
+
+- Only includes citations registered in Crossref metadata
+- Reference parsing depends on publisher-provided metadata quality
+- Some arXiv works may be cited by their published version DOI rather than arXiv DOI
+- Does not include citations from works not registered with Crossref
+
+## License
+
+This dataset is released under [CC0 1.0 Universal](https://creativecommons.org/publicdomain/zero/1.0/). The underlying Crossref metadata is available under similar open terms.
+
+## Citation
+
+If you use this dataset, please cite:
+
+```bibtex
+@dataset{crossref_arxiv_citations_2025,
+  title = {Crossref arXiv Citations},
+  author = {Cometadata},
+  year = {2025},
+  publisher = {Hugging Face},
+  url = {https://huggingface.co/datasets/cometadata/crossref-arxiv-citations}
+}
+```

data/arxiv_citations_valid.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e28c272443db9050bc9595037ea1bb09adf822719fcdf7abd2a4e449a4ff5047
+size 1865860050