Machine Translation on FLORES-200 (xho_Latn->en, test)
82.1COMET-22SCALE-refine
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| SCALE-refineLLM=GPT-4, STM=NLLB-3.3B2023.09 | 82.1 | 67.3 | 75.7 | 40 | |
| NLLBModel Version=3.3B, Evaluation Protocol=Supervised2023.09 | 80.7 | 65.8 | 74 | 40.1 | |
| GPT-4Evaluation Protocol=few-shot2023.09 | 78.8 | 67.1 | 70.8 | 34.5 | |
| GPT-3.5Evaluation Protocol=few-shot2023.09 | 69.1 | 65.5 | 58.3 | 21.9 | |
| M2M100Type=Specialized Model2023.09 | 68 | 62.1 | 59 | 25.7 | |
| XGLMEvaluation Protocol=few-shot2023.09 | 39.6 | 41.7 | 37.1 | 1.6 |