The SourceData-NLP dataset: integrating curation into scientific publishing for training large language models

Open in new window