This dataset contains pairs of extracted text and their corresponding headwords from all editions of Nordisk Familjebok. Each entry includes a text and a headword field. If a text has a corresponding headword, the headword field is populated; otherwise, it contains an empty string. This dataset is designed for training and evaluating headword extraction models.
