FMMD: A multimodal multidisciplinary dataset of open peer reviews from F1000Research

FMMD is a multimodal and multidisciplinary open peer review dataset constructed from F1000Research, designed to support research on Automated Scholarly Paper Review (ASPR) and multimodal scientific document understanding. The dataset preserves different versions of F1000Research articles, original HTML pages, image resources, and structured open peer review information, including review outcomes, reviewer comments, and author responses. FMMD is released in two parts: NativeFMMD and ProcessedFMMD. The current release contains 8,939 scholarly articles, 26,123 reviewer reports, 8,887 author responses, and 49,082 multimodal items. NativeFMMD: NativeFMMD_full_versions_url_v1.0.1.zip: Contains the F1000Research page URLs corresponding to the articles and their different versions. NativeFMMD_articles_htmls_valid_v1.0.1.zip: Contains the original HTML files for all valid article pages. ProcessedFMMD: ProcessedFMMD_part1_1-9_v1.0.1.7z, ProcessedFMMD_part2_10-11_v1.0.1.7z, ProcessedFMMD_part3_12-14_v1.0.1.7z: Contain data files for different article versions, including the original HTML pages, cleaned HTML files with tags removed (view.html), and the corresponding image resources. Due to the large data volume, the files are divided into three compressed archives. ProcessedFMMD_articles_valid_v1.0.1.csv: Contains structured open peer review data extracted from the HTML pages of different F1000Research article versions. ProcessedFMMD_multimodal_extracted_v1.0.1.csv: Contains multimodal-related reviewer–author interaction sentence pairs extracted from open peer review records. ProcessedFMMD_collections_v1.0.1.csv: Contains structured metadata for 49 F1000Research thematic Collections. ProcessedFMMD_gateways_data_v1.0.1.csv: Contains structured metadata for F1000Research Gateways. Related publication:Zhuang, Z., Fu, Y., Zhu, J., Zhou, Z., & Lin, J. (2026). FMMD: A multimodal multidisciplinary dataset of open peer reviews from F1000Research. arXiv. https://doi.org/10.48550/arXiv.2602.14285