Machine Translation on FLORES-200 asm_Beng-English (test)
86.6COMET-22SCALE-refine
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| SCALE-refineLLM=GPT-4, STM=NLLB-3.3B2023.09 | 86.6 | 83.2 | 73.8 | 34.1 | |
| NLLBModel Version=3.3B, Evaluation Protocol=Supervised2023.09 | 85.6 | 82.8 | 72.1 | 33.9 | |
| GPT-4Evaluation Protocol=few-shot2023.09 | 83.9 | 80.9 | 69.1 | 27.9 | |
| MicrosoftType=Commercial System2023.09 | 83.5 | 81.7 | 68.8 | 29.6 | |
| GPT-3.5Evaluation Protocol=few-shot2023.09 | 78.6 | 76.7 | 61 | 18.1 | |
| XGLMEvaluation Protocol=few-shot2023.09 | 62.7 | 57.8 | 38.8 | 3.7 |