Semantic-conditioned sentence generation on FrameNet Before Filtering 1.7 (test)
0.979FE Fid.Human (FN 1.7)
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Human (FN 1.7)2024.06 | 0.979 | 78.1 | 1 | |
| Pancholy et al.2024.06 | 0.953 | 127.8 | 0.611 | |
| T5 | Frame + FEConditioning=Frame+FE-Conditioning2024.06 | 0.882 | 136.8 | 0.644 | |
| T5 | FEConditioning=FE-Conditioning2024.06 | 0.862 | 127.6 | 0.711 | |
| GPT-4 | Frame + FEConditioning=Frame+FE-Conditioning2024.06 | 0.853 | 117.2 | 0.733 | |
| GPT-4 | FEConditioning=FE-Conditioning2024.06 | 0.841 | 106.3 | 0.7 | |
| T5Conditioning=None2024.06 | 0.784 | 139.3 | 0.594 | |
| GPT-4Conditioning=None2024.06 | 0.704 | 114.9 | 0.528 |