KNoTE: A TEI Encoded XML Dataset of Modern Korean Novels

KNoTE (Korean Novel TEI Encoded) is an open XML dataset of 33 modern Korean novels (1906–1954) by 22 authors, encoded according to the TEI P5 Guidelines. Source texts were collected from Korean Wikisource and semi-automatically encoded using a large language model via prompt engineering, followed by multi-layered manual validation by seven annotators. The markup captures named entities, character ontologies, direct speech, Hanja annotations, and period-specific metadata including political ideologies, colonial-era place names, and social class markers. As the first TEI-conformant corpus of modern Korean literature, the dataset supports computational literary analysis, NLP model fine-tuning, and cross-corpus interoperability.