NLG Meta-evaluation on WMT en-cs 24
0.318Kendall CorrelationLlama 4 Scout
Evaluation Results
| Method | Links | |
|---|---|---|
| Llama 4 ScoutShot=Zero2026.03 | 0.318 | |
| Llama 3.3 70BShot=Zero2026.03 | 0.307 | |
| Llama 3.3 70BShot=Few2026.03 | 0.238 | |
| Qwen 3 30BShot=Few2026.03 | 0.153 | |
| Llama 4 ScoutShot=Few2026.03 | 0.148 | |
| Qwen 3 30BShot=Zero2026.03 | 0.085 |