Machine Translation on WMT En-De 2023 (test)
84.3COMET ScoreDUAL-REFLECT
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| DUAL-REFLECTRefinement=Dual-Reflection2024.06 | 84.3 | 71 | |
| ChatGPT + MADRefinement=Multi-Agent Debate2024.06 | 83.9 | 70.3 | |
| ChatGPT + 5shotPrompting=5-shot2024.06 | 83.7 | 69.4 | |
| ChatGPTPrompting=Zero-shot2024.06 | 83.5 | 69.1 |