LEGAR BENCH is the first large-scale Korean legal case retrieval benchmark, comprising two dataset versions tailored to different evaluation needs:
from datasets import load_dataset
standard = load_dataset("Chaeeun-Kim/LEGAR_BENCH", "standard")
stricter = load_dataset("Chaeeun-Kim/LEGAR_BENCH", "stricter")
stricter_by_difficulty = load_dataset("Chaeeun-Kim/LEGAR_BENCH", "stricter_by_difficulty")
corpus = load_dataset("Chaeeun-Kim/LEGAR_BENCH", "corpus")… See the full description on the dataset page:
https://huggingface.co/datasets/Chaeeun-Kim/LEGAR_BENCH.
