Commonsense Reasoning on HellaSwag (leaderboard)
95.6AccuracyHuman
Evaluation results
| Method | Links | |
|---|---|---|
| Human2021.03 | 95.6 | |
| UNICORN2021.03 | 93.9 | |
| RoBERTa-Large Ensembleensemble=true2021.03 | 85.5 | |
| HYKAS+CSKG2021.03 | 85 | |
| RoBERTa-Large2021.03 | 81.7 |
| Method | Links | |
|---|---|---|
| Human2021.03 | 95.6 | |
| UNICORN2021.03 | 93.9 | |
| RoBERTa-Large Ensembleensemble=true2021.03 | 85.5 | |
| HYKAS+CSKG2021.03 | 85 | |
| RoBERTa-Large2021.03 | 81.7 |