Nile Translation on Nile
72Pairwise AccuracyPara-PLUIE
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Para-PLUIEJudge Model=Llama2026.02 | 72 | 46 | 43 | |
| *-PLUIEJudge Model=Llama2026.02 | 70 | 43 | 42 | |
| Para-PLUIEJudge Model=Phi2026.02 | 70 | 44 | 42 | |
| *-PLUIEJudge Model=Phi2026.02 | 69 | 42 | 40 | |
| LLM-LikertJudge Model=Llama2026.02 | 69 | 40 | 44 | |
| BERTScore2026.02 | 68 | 43 | 40 | |
| METEOR2026.02 | 68 | 40 | 37 | |
| LLM-LikertJudge Model=Phi2026.02 | 65 | 37 | 41 | |
| Levenshtein2026.02 | 64 | 34 | 29 | |
| Modern BERTScore2026.02 | 63 | 41 | 36 | |
| LLM-choiceJudge Model=Phi2026.02 | 47 | 42 | 20 | |
| LLM-choiceJudge Model=Llama2026.02 | 47 | 43 | 20 | |
| BLEU2026.02 | 43 | — | 0 | |
| Random weighted2026.02 | 43 | 6 | 0 | |
| Random uniform2026.02 | 33 | 6 | 0 |