Webis-Dataset-Reviews-21

The Webis-Dataset-Reviews-21 corpus comprises the curated list of 13,372 NLP-related datasets and their 539,411 mentions extracted from all the publications available in ACL Anthology corpus. <strong>Dataset specification</strong> All files are in gzip-compressed JSON Lines format. <strong>dataset_mentions.jsonl:</strong> contains the extracted dataset mentions <pre><code class="language-json">[dataset, paper, mention]</code></pre> <strong>nlp_datasets.jsonl:</strong> each record contains the following dataset metadata <pre><code class="language-json">[source, name, doi, decsription, year, creator, corpus_url, paper_title, paper_url, task, language, format, size, ids]</code></pre>