This dataset contains 9.21K rows of Arabic instruction-based reasoning QA pairs. It is a comprehensive collection of data points, meticulously crafted to enhance Arabic language reasoning capabilities for models.
The dataset was generated by combining:
♦︎ Data from the Hugging Face dataset MohammedNasser/Arabic_Reasoning_Instruct_QA. | Size = 1.23K instructions
♦︎ Synthetic data generated using the GPT-4o-Mini API… See the full description on the dataset page: https://huggingface.co/datasets/Omartificial-Intelligence-Space/Arabic_Reasoning_Dataset.
