Machine Translation on MT (test)
58.24Average Win RateALARM
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| ALARMEvaluator=Grammar Error Rate2024.03 | 58.24 | 57.8 | 57.53 | 59.37 | — | |
| ALARMEvaluator=Holistic Reward2024.03 | 52.12 | 51.7 | 52.93 | 51.74 | — | |
| ALARMEvaluator=gpt-3.5-turbo2024.03 | 51.73 | 51.03 | 52.15 | 52.02 | — | |
| Holistic RewardEvaluator=gpt-3.5-turbo2024.03 | 50.43 | — | 51.03 | 51.27 | 48.97 | |
| Holistic RewardEvaluator=Holistic Reward2024.03 | 50 | — | 51.52 | 50.18 | 48.3 | |
| Weighted SumEvaluator=Holistic Reward2024.03 | 49.8 | 49.82 | 51.31 | — | 48.26 | |
| Weighted SumEvaluator=gpt-3.5-turbo2024.03 | 49.24 | 48.73 | 51.02 | — | 47.98 | |
| Holistic RewardEvaluator=Grammar Error Rate2024.03 | 49.19 | — | 51.9 | 53.47 | 42.2 | |
| Grammar RewardEvaluator=gpt-3.5-turbo2024.03 | 48.6 | 48.97 | — | 48.98 | 47.85 | |
| Grammar RewardEvaluator=Holistic Reward2024.03 | 48.08 | 48.47 | — | 48.69 | 47.07 | |
| Grammar RewardEvaluator=Grammar Error Rate2024.03 | 47.07 | 48.1 | — | 50.63 | 42.47 | |
| Weighted SumEvaluator=Grammar Error Rate2024.03 | 45.51 | 46.54 | 49.37 | — | 40.63 |