{
  "dataset": "multi-task-dataset",
  "version": "v1.1",
  "total_examples": 100000001,
  "splits": {
    "train": 98000000,
    "validation": 1000000,
    "test": 1000001
  },
  "format": "parquet",
  "compression": "zstd",
  "compression_level": 3,
  "seed": 20260916,
  "batch_size": 100000,
  "task_families": [
    "math",
    "logic",
    "code",
    "research",
    "fact_check",
    "data_analysis",
    "reasoning",
    "self_correction",
    "instruction_following",
    "planning",
    "adversarial",
    "calibration",
    "ambiguity",
    "counterexample",
    "constraint_reasoning",
    "error_analysis",
    "generalization",
    "prompt_review",
    "consistency_check",
    "evidence_check"
  ],
  "synthetic": true
}