English International Linguistics Olympiad problem dataset built from the official
IOL problem and solution PDFs at
https://ioling.org/problems/by_year/.
This release contains a provenance-rich full dataset, a text-extraction strict
split, manual verification reports, and deterministic scoring utilities.
train.parquet: text-extraction strict split, flattened for
datasets.load_dataset("agurung/ioling"). This split is useful for
evaluation experiments… See the full description on the dataset page:
https://huggingface.co/datasets/agurung/ioling.
