Description unarXive is a scholarly data set containing publications' full-text, annotated in-text citations, and a citation network. The data is generated from all LaTeX sources on arXiv and therefore of higher quality than data generated from PDF files. Typical use cases are Citation recommendation Citation context analysis Bibliographic analyses Reference string parsing This version (v3) of our data set is based on all arXiv publications until 2020-07-31 and on the Microsoft Academic Graph as of 2020-08-18. As additional contribution, we included a table with the publication date and the scientific discipline for each paper for easier filtering. Note: This Zenodo record is an old version of unarXive. You can find the most recent version at https://zenodo.org/record/7752754 and https://zenodo.org/record/7752615 Access ┏━━━━━━━━━━━━━━━━━━━━━━━━━━┓┃ D O W N L O A D S A M P L E ┃┗━━━━━━━━━━━━━━━━━━━━━━━━━━┛ To download the whole data set send an access request and note the following: Note: this Zenodo record is a "full" version of unarXive, which was generated from all of arXiv.org including non-permissively licensed papers. Make sure that your use of the data is compliant with the paper's licensing terms.¹ ¹ For information on papers' licenses use arXiv's bulk metadata access. The code used for generating the data set is publicly available. Usage examples for our data set are provided at here on GitHub. Citing This initial version of unarXive is described in the following journal article. Tarek Saier, Michael Färber: "unarXive: A Large Scholarly Data Set with Publications' Full-Text, Annotated In-Text Citations, and Links to Metadata", Scientometrics, 2020,[link to an author copy] The updated version is described in the following conference paper. Tarek Saier, Michael Färber. "unarXive 2022: All arXiv Publications Pre-Processed for NLP, Including Structured Full-Text and Citation Network", JCDL 2023.[link to an author copy]