Document-level Machine Translation on Document-level Translation Averaged
67.12Coherence ScoreOurs
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| OursBase Model=LLaMA-3-8B-Instruct, Quality-aware Fine-tuning=true2025.04 | 67.12 | 43.53 | 66.57 | 26.51 | 59.86 | |
| OursBase Model=LLaMA-3-8B-Instruct, Quality-aware Fine-tuning=false2025.04 | 66.07 | 43.06 | 65.98 | 31.64 | 59.57 | |
| DocRefineBase Model=LLaMA-3-8B-Instruct, Refinement Level=document2025.04 | 65.09 | 42.8 | 63.68 | 31.62 | 59.01 | |
| DocRefineBase Model=LLaMA-3-8B-Instruct, Refinement Level=sentence2025.04 | 64.95 | 43 | 63.62 | 30.13 | 58.69 | |
| SentRefineBase Model=LLaMA-3-8B-Instruct, Refinement Level=sentence2025.04 | 64.27 | 43.09 | 60.08 | 32.14 | 57.47 | |
| Doc2DocBase Model=LLaMA-3-8B-Instruct, Fine-tuned=true2025.04 | 63.42 | 42.99 | 64.99 | 31.58 | 57.86 | |
| Doc2DocBase Model=LLaMA-3-8B-Instruct2025.04 | 62.28 | 40.04 | 61.25 | 31.85 | 51.3 | |
| Sent2SentBase Model=LLaMA-3-8B-Instruct, Fine-tuned=true2025.04 | 56.23 | 42.94 | 60.45 | 30.34 | 58.61 | |
| Sent2SentBase Model=LLaMA-3-8B-Instruct2025.04 | 56.17 | 42.57 | 57.23 | 32.86 | 48.49 |