Machine Translation on FLORES-200 npi_Deva-English (test)
91.1COMET-22SCALE-refine
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| SCALE-refineLLM=GPT-4, STM=NLLB-3.3B2023.09 | 91.1 | 88.8 | 78.1 | 44 | |
| NLLBModel Version=3.3B, Evaluation Protocol=Supervised2023.09 | 90.4 | 88.3 | 77.1 | 45 | |
| GPT-4Evaluation Protocol=few-shot2023.09 | 90.2 | 88.1 | 76.3 | 40.8 | |
| MicrosoftType=Commercial System2023.09 | 89.8 | 88.2 | 75.3 | 42.8 | |
| GPT-3.5Evaluation Protocol=few-shot2023.09 | 87.2 | 85.4 | 69.9 | 29.3 | |
| M2M100Type=Specialized Model2023.09 | 75.2 | 73.6 | 55.1 | 21.2 | |
| XGLMEvaluation Protocol=few-shot2023.09 | 72.9 | 67 | 48.8 | 8.3 |