Question-Answer Generation on FairytaleQA (val and test)
3.03Diversity (Q)FQAG
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| FQAGevaluation_mode=Human evaluation (global ranking and local scoring)2023.06 | 3.03 | 3.06 | 2.66 | 2.65 | 2.14 | 1.74 | 2.64 | 1.11 | |
| SQGevaluation_mode=Human evaluation (global ranking and local scoring)2023.06 | 2.96 | 3.03 | 3.3 | 2.44 | 1.87 | 1.34 | 2.55 | 1.36 | |
| QAgenbackbone=BART-large, ranking_model=RoBERTa-base, evaluation_mode=Human evaluation (global ranking and local scoring)2023.06 | 2.35 | 2.18 | 2.35 | 2.69 | 2.22 | 1.9 | 2.35 | 1.98 | |
| GTtype=Ground Truth, evaluation_mode=Human evaluation (global ranking and local scoring)2023.06 | 1.65 | 1.71 | 1.68 | 2.97 | 2.65 | 2.5 | 2.8 | 1.95 |