{
  "id": "ds_mmlu_v1",
  "name": "MMLU",
  "description": "Measuring Massive Multitask Language Understanding — multiple-choice benchmark spanning 57 subjects from STEM and international law to nutrition and religion.",
  "version": "1.0.0",
  "task": "knowledge",
  "source": "https://huggingface.co/datasets/cais/mmlu",
  "license": "MIT",
  "splits": {
    "dev": { "count": 285 },
    "validation": { "count": 1531 },
    "test": { "count": 14042 }
  },
  "case_count": 15908,
  "tags": ["knowledge", "multitask", "multiple-choice", "benchmark"],
  "created": "2020-09-07T00:00:00Z",
  "modified": "2026-05-22T00:00:00Z"
}