Table 10.
Template selection average accuracies per language. The models are different for each language and also the number of candidates per task. Results for all tasks can be found in Table A2 of Appendix B. Bold values indicate the best results for each language.
Accuracy | N-Grams (N = 3) | GPT-2 1 Epoch | GPT-2 2 Epochs |
---|---|---|---|
es | 38.85 | 52.23 | 84.35 |
fr | 39.75 | 49.68 | 62.39 |
no | 26.47 | 76.94 | 73.86 |