jakequist/kanjiland-silver-annotations

Kanjiland — Silver Annotations 9,388 Japanese sentences annotated in the full Kanjiland reading-comprehension format: morpheme segmentation, furigana (ruby on kanji runs), per-token contextual glosses, word groupings, an English sentence translation, and grammar-pattern labels from a closed 120-rule inventory. Part of Kanjiland, a from-scratch Japanese reading-comprehension engine. Format Each line of silver_annotations.jsonl is: {"ja": "<Japanese sentence>"… See the full description on the dataset page: https://huggingface.co/datasets/jakequist/kanjiland-silver-annotations.