Machine Translation on En-Es document-level
87.31d-COMETOurs
Evaluation results
| Method | Links | |
|---|---|---|
| OursBase Model=Mistral-Nemo-Instruct, Quality-aware fine-tuning=true2025.04 | 87.31 | |
| DocRefine (doc)Base Model=Mistral-Nemo-Instruct2025.04 | 86.95 | |
| Ours (-QA)Base Model=Mistral-Nemo-Instruct, Quality-aware fine-tuning=false2025.04 | 86.94 | |
| DocRefine (sent)Base Model=Mistral-Nemo-Instruct2025.04 | 86.91 | |
| OursBase Model=LLaMA-3-8B-Instruct, Quality-aware fine-tuning=true2025.04 | 86.86 | |
| Doc2Doc (tuned)Base Model=Mistral-Nemo-Instruct2025.04 | 86.72 | |
| Ours (-QA)Base Model=LLaMA-3-8B-Instruct, Quality-aware fine-tuning=false2025.04 | 86.6 | |
| DocRefine (doc)Base Model=LLaMA-3-8B-Instruct2025.04 | 86.49 | |
| SentRefine (sent)Base Model=Mistral-Nemo-Instruct2025.04 | 86.48 | |
| SentRefine (sent)Base Model=LLaMA-3-8B-Instruct2025.04 | 86.34 | |
| DocRefine (sent)Base Model=LLaMA-3-8B-Instruct2025.04 | 86.32 | |
| Sent2Sent (tuned)Base Model=LLaMA-3-8B-Instruct2025.04 | 86.11 | |
| CPLBase Model=Qwen3-8B, Input Type=Context-aware2026.03 | 86.05 | |
| CPLBase Model=Llama-3-8B-Instruct, Input Type=Context-aware2026.03 | 86.04 | |
| CPLBase Model=Qwen3-8B, Input Type=Sentence-level2026.03 | 86 | |
| CPLBase Model=Llama-3-8B-Instruct, Input Type=Sentence-level2026.03 | 85.98 | |
| CrossLossBase Model=Qwen3-8B, Input Type=Context-aware2026.03 | 85.91 | |
| IntraLossBase Model=Qwen3-8B, Input Type=Context-aware2026.03 | 85.88 | |
| IntraLossBase Model=Qwen3-8B, Input Type=Sentence-level2026.03 | 85.85 | |
| IntraLossBase Model=Llama-3-8B-Instruct, Input Type=Context-aware2026.03 | 85.84 | |
| CrossLossBase Model=Qwen3-8B, Input Type=Sentence-level2026.03 | 85.83 | |
| Ctx-Aware_CPOBase Model=Qwen3-8B, Input Type=Context-aware2026.03 | 85.79 | |
| IntraLossBase Model=Llama-3-8B-Instruct, Input Type=Sentence-level2026.03 | 85.79 | |
| Sent2Sent (tuned)Base Model=Mistral-Nemo-Instruct2025.04 | 85.77 | |
| CrossLossBase Model=Llama-3-8B-Instruct, Input Type=Context-aware2026.03 | 85.71 | |
| Sent-Level_CPOBase Model=Qwen3-8B, Input Type=Sentence-level2026.03 | 85.69 | |
| Ctx-Aware_CPOBase Model=Llama-3-8B-Instruct, Input Type=Context-aware2026.03 | 85.67 | |
| CrossLossBase Model=Llama-3-8B-Instruct, Input Type=Sentence-level2026.03 | 85.65 | |
| Sent-Level_CPOBase Model=Llama-3-8B-Instruct, Input Type=Sentence-level2026.03 | 85.63 | |
| Doc2Doc (tuned)Base Model=LLaMA-3-8B-Instruct2025.04 | 85.61 | |
| Sent2SentBase Model=Mistral-Nemo-Instruct2025.04 | 85.47 | |
| Sent+Ctx_tuned+Base Model=Llama-3-8B-Instruct, Input Type=Context-aware2026.03 | 85.43 | |
| Sent+Ctx_tuned+Base Model=Qwen3-8B, Input Type=Context-aware2026.03 | 85.41 | |
| Sent+Ctx_tuned+Base Model=Llama-3-8B-Instruct, Input Type=Sentence-level2026.03 | 85.39 | |
| CPLBase Model=Qwen3-4B, Input Type=Context-aware2026.03 | 85.38 | |
| Sent+Ctx_tunedBase Model=Llama-3-8B-Instruct, Input Type=Context-aware2026.03 | 85.38 | |
| Doc2DocBase Model=LLaMA-3-8B-Instruct2025.04 | 85.36 | |
| Sent+Ctx_tunedBase Model=Qwen3-8B, Input Type=Context-aware2026.03 | 85.35 | |
| Sent+Ctx_tunedBase Model=Llama-3-8B-Instruct, Input Type=Sentence-level2026.03 | 85.34 | |
| Sent+Ctx_tunedBase Model=Qwen3-8B, Input Type=Sentence-level2026.03 | 85.27 | |
| Ctx-Aware_tunedBase Model=Qwen3-8B, Input Type=Context-aware2026.03 | 85.25 | |
| CrossLossBase Model=Qwen3-4B, Input Type=Context-aware2026.03 | 85.24 | |
| Sent+Ctx_tuned+Base Model=Qwen3-8B, Input Type=Sentence-level2026.03 | 85.24 | |
| CPLBase Model=Qwen3-4B, Input Type=Sentence-level2026.03 | 85.22 | |
| Ctx-Aware_tunedBase Model=Llama-3-8B-Instruct, Input Type=Context-aware2026.03 | 85.2 | |
| Sent-Level_tunedBase Model=Qwen3-8B, Input Type=Sentence-level2026.03 | 85.17 | |
| Ctx-AwareBase Model=Qwen3-8B, Input Type=Context-aware2026.03 | 85.15 | |
| IntraLossBase Model=Qwen3-4B, Input Type=Context-aware2026.03 | 85.11 | |
| IntraLossBase Model=Qwen3-4B, Input Type=Sentence-level2026.03 | 85.1 | |
| CrossLossBase Model=Qwen3-4B, Input Type=Sentence-level2026.03 | 85.07 | |
| Sent-Level_tunedBase Model=Llama-3-8B-Instruct, Input Type=Sentence-level2026.03 | 85.07 | |
| Ctx-Aware_CPOBase Model=Qwen3-4B, Input Type=Context-aware2026.03 | 85.03 | |
| Sent2SentBase Model=LLaMA-3-8B-Instruct2025.04 | 85.02 | |
| Sent-LevelBase Model=Qwen3-8B, Input Type=Sentence-level2026.03 | 85.01 | |
| Sent-Level_CPOBase Model=Qwen3-4B, Input Type=Sentence-level2026.03 | 84.9 | |
| Sent+Ctx_tuned+Base Model=Qwen3-4B, Input Type=Context-aware2026.03 | 84.55 | |
| Sent+Ctx_tuned+Base Model=Qwen3-4B, Input Type=Sentence-level2026.03 | 84.49 | |
| Ctx-AwareBase Model=Qwen3-4B, Input Type=Context-aware2026.03 | 84.46 | |
| Sent+Ctx_tunedBase Model=Qwen3-4B, Input Type=Context-aware2026.03 | 84.46 | |
| Sent+Ctx_tunedBase Model=Qwen3-4B, Input Type=Sentence-level2026.03 | 84.44 | |
| Doc2DocBase Model=Mistral-Nemo-Instruct2025.04 | 84.34 | |
| Ctx-Aware_tunedBase Model=Qwen3-4B, Input Type=Context-aware2026.03 | 84.26 | |
| Sent-Level_tunedBase Model=Qwen3-4B, Input Type=Sentence-level2026.03 | 84.25 | |
| Sent-LevelBase Model=Qwen3-4B, Input Type=Sentence-level2026.03 | 81.29 | |
| Sent-LevelBase Model=Llama-3-8B-Instruct, Input Type=Sentence-level2026.03 | 80.56 | |
| Ctx-AwareBase Model=Llama-3-8B-Instruct, Input Type=Context-aware2026.03 | 50.38 |