1{2 "document_id": "lb_qasper_6b9d0c1e63714067fc0b54c024dc065e63f64cbaf44ab4d0",3 "family": "public_qasper",4 "split": "test",5 "document_hash": "1499a1624a4f52a3ba09cc0d27d857c0254b49316cb332fa18a46e022b0ed127",6 "questions": [7 {8 "question_id": "q00",9 "qtype": "qasper",10 "question": "by how much did their model improve?",11 "answer": "For the WikiLarge dataset, the improvement over baseline NMT is 2.11 BLEU, 1.7 FKGL and 1.07 SARI.\nFor the WikiSmall dataset, the improvement over baseline NMT is 8.37 BLEU.",12 "aliases": [13 "6.37 BLEU"14 ],15 "answer_type": "qa_f1",16 "tolerance": 0.0,17 "supporting_fact_ids": [],18 "answer_position_frac": -1.019 }20 ],21 "facts": []22}