Knowledge-grounded Dialog Generation on WoW (Seen)
4.5Appropriateness ScoreHuman Response
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Human ResponseEvaluation Set=DiffKS Comparison2022.05 | 4.5 | 4.3 | |
| Human ResponseEvaluation Set=Base vs KI2022.05 | 4.4 | 4.3 | |
| DiffKS+KI2022.05 | 3.9 | 4 | |
| Transformer+KI2022.05 | 3.7 | 3.5 | |
| DiffKS2022.05 | 3.6 | 3.6 | |
| Transformer2022.05 | 3 | 3.2 |