MizanQA is the first benchmark designed to evaluate large language models (LLMs) on Moroccan legal question answering tasks.It contains 1,776 multiple-choice questions (MCQs), many with multiple correct answers, covering diverse areas of Moroccan law.The dataset reflects the linguistic and legal complexity of Morocco, where statutes are influenced by Modern Standard Arabic, Maliki jurisprudence… See the full description on the dataset page: https://huggingface.co/datasets/adlbh/MizanQA-v0.
