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.
Each line of silver_annotations.jsonl is:
{"ja": "
"… See the full description on the dataset page: https://huggingface.co/datasets/jakequist/kanjiland-silver-annotations.
