{
  "document_id": "lb_qasper_4eb3ad14a5fbb16b03f54ef42febd2f167f8b2a9f5a26343",
  "family": "public_qasper",
  "split": "test",
  "document_hash": "9bbabd24a8439aaea69e00c1a9d6d68e70a4a4dc90d8c63e69bcd733d3664777",
  "questions": [
    {
      "question_id": "q00",
      "qtype": "qasper",
      "question": "What is the combination of rewards for reinforcement learning?",
      "answer": "irony accuracy, sentiment preservation",
      "aliases": [
        " irony accuracy and sentiment preservation"
      ],
      "answer_type": "qa_f1",
      "tolerance": 0.0,
      "supporting_fact_ids": [],
      "answer_position_frac": -1.0
    }
  ],
  "facts": []
}