Machine Translation on WMT22 En-Ja (news)
89.5COMETIBUT
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| IBUTMode=Iterative Bilingual Understanding (IBUT)2024.10 | 89.5 | 69.9 | 24.3 | |
| ChatGPTMode=Iterative Refinement2024.10 | 88.6 | 67.9 | 23.4 | |
| ChatGPTMode=Multi-Agent Prompt Selection (MAPS)2024.10 | 88.5 | 67.4 | 21.2 | |
| ChatGPTMode=Multi-Agent Discussion (MAD)2024.10 | 88.4 | 67.9 | 19.7 | |
| ChatGPTMode=Zero-shot2024.10 | 88.3 | 66.3 | 17.3 | |
| ChatGPTMode=Dual-Reflection2024.10 | 88.3 | 67.2 | 20.2 | |
| ChatGPTMode=5-shot prompting2024.10 | 88.2 | 67.1 | 17.9 | |
| ChatGPTMode=Reranking2024.10 | 88 | 66.6 | 21.2 | |
| ChatGPTMode=Translation Error Analysis and Reduction (TEaR)2024.10 | 88 | 67.3 | 23.4 |