This dataset contains Chinese healthcare question-answer pairs from the AIR-Bench 24.05 benchmark, specifically focusing on the default development and test splits. The data is formatted as prompt-completion pairs where prompts represent medical queries and completions reference specific document IDs. It is designed for evaluating… See the full description on the dataset page: https://huggingface.co/datasets/Roshang09112007/adaption-air-bench-healthcare-zh.