Question Generation on Fairytale QA
54.8ROUGE-LM
Evaluation Results
| Method | Links | |
|---|---|---|
| Mmode=upperbound2022.09 | 54.8 | |
| BARTtraining=fine-tuned, author=Xu et al., 20222022.09 | 52.7 | |
| APS + round-triptype=ensemble2022.09 | 43.9 | |
| tri-gram + APS + round-triptype=ensemble2022.09 | 43.5 | |
| round-tripselection_criteria=QA semantic equivalence2022.09 | 43.4 | |
| bi-gram + APS + round-triptype=ensemble2022.09 | 43.1 | |
| tri-gram + round-triptype=ensemble2022.09 | 43 | |
| bi-gram + round-triptype=ensemble2022.09 | 42.9 | |
| Mgmode=greedy2022.09 | 42.4 | |
| tri-gram + APStype=ensemble2022.09 | 40.9 | |
| averaged prompt score (APS)selection_criteria=Prompt-based Score2022.09 | 40.6 | |
| bi-gram + APStype=ensemble2022.09 | 40.6 | |
| bi-gramselection_criteria=n-gram similarity, n=22022.09 | 40.3 | |
| tri-gramselection_criteria=n-gram similarity, n=32022.09 | 40.3 | |
| Msmode=sample avg2022.09 | 39.9 | |
| overall prompt score (OPS)selection_criteria=Prompt-based Score2022.09 | 39.9 | |
| Msmode=lowerbound2022.09 | 25.9 |