Human Evaluation of Multilingual Capabilities on Multilingual human evaluation
4.15Score (Arabic)Claude-3-Opus
Evaluation Results
| Method | Links | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Claude-3-OpusVersion=02292024.07 | 4.15 | 4.23 | 4.4 | 3.85 | 4.23 | 4.09 | 4.25 | 4.31 | 4.01 | 3.98 | 4.15 | |
| Qwen2-72B-Instruct2024.07 | 3.86 | 4.01 | 3.83 | 3.63 | 4.14 | 3.97 | 4.15 | 4.1 | 3.75 | 3.91 | 3.93 | |
| GPT-4oVersion=05132024.07 | 3.55 | 4.16 | 4.39 | 3.72 | 4.4 | 3.89 | 4.32 | 4.26 | 4.09 | 4.14 | 4.09 | |
| GPT-4-TurboVersion=04092024.07 | 3.44 | 4.19 | 4.09 | 3.68 | 4.24 | 3.86 | 4.27 | 4.08 | 4.11 | 3.84 | 3.98 | |
| GPT-3.5-TurboVersion=11062024.07 | 2.52 | 3.47 | 3.56 | 2.75 | 2.37 | 3.37 | 3.24 | 4.07 | 3.38 | 3.9 | 3.16 |