Machine Translation on RefuteBench Machine Translation 1.0 (test)
87.6FAClaude-2
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Claude-22024.02 | 87.6 | 72.19 | 31.23 | 62.65 | 18.37 | 50.31 | |
| ChatGPT2024.02 | 66.8 | 15.7 | 9.9 | 33.4 | 10.9 | 26.44 | |
| GPT-42024.02 | 65.6 | 56 | 52.1 | 46.6 | 33.1 | 69.07 | |
| LLAMA-2-13B-ChatParameters=13B, Model Type=Chat2024.02 | 59.2 | 71.48 | 12.12 | 40.75 | 5.01 | 16.21 | |
| LLAMA-2-7B-ChatParameters=7B, Model Type=Chat2024.02 | 45.2 | 65.2 | 5.2 | 30.7 | 3.4 | 11.38 | |
| ALPACA-7BParameters=7B2024.02 | 16.8 | 29.4 | 10.2 | 36.9 | 18.6 | 13.86 | |
| Mistral-7B-Instruct-v0.2Parameters=7B, Model Type=Instruct2024.02 | 14 | 44.2 | 20 | 24.9 | 12.6 | 27.01 |