This dataset is a Hungarian translation/alignment of the MMLU (Massive Multitask Language Understanding) benchmark, organized per subject (one Hugging Face config per subject), mirroring the subject structure of the Hub-hosted MMLU distribution.
Upstream reference dataset: cais/mmluLicense: MIT (kept consistent with the upstream distribution)
⚠️ Translation note: This is a translated benchmark. Residual artifacts (formatting, terminology drift, or occasional awkward… See the full description on the dataset page:
https://huggingface.co/datasets/NYTK/hu-mmlu.
