Simultaneous Machine Translation on ACL60/60 En-De (eval)
100BLEUReference
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| ReferenceTuning method=-, Tuning data=-, Inference prompt=-2026.01 | 100 | 100 | 0 | 0.9511 | 0.8048 | 2.507 | |
| ReferenceTuning method=–, Tuning data=–, Inference prompt=–2026.01 | 100 | 100 | 0 | 0.9511 | 0.8048 | 2.507 | |
| Step-wise (Drop / Cut)Tuning method=-, Tuning data=-, Inference prompt=Incremental drop / cut selection2026.01 | 49.97 | 70.96 | 37.38 | 0.8594 | 0.8111 | 1.18 | |
| Step-wise (Drop / Cut)Tuning method=–, Tuning data=–, Inference prompt=Incremental drop / cut selection2026.01 | 49.97 | 70.96 | 37.38 | 0.8594 | 0.8111 | 1.18 | |
| Step-wise (Actions)Tuning method=-, Tuning data=-, Inference prompt=Incremental action selection2026.01 | 47.8 | 70.08 | 38.72 | 0.8541 | 0.8108 | 1.173 | |
| Step-wise (Actions)Tuning method=–, Tuning data=–, Inference prompt=Incremental action selection2026.01 | 47.8 | 70.08 | 38.72 | 0.8541 | 0.8108 | 1.173 | |
| SalamiTuning method=Fine-tuning, Tuning data=Dev set (Salami refs), Inference prompt=Salami prompt2026.01 | 47.48 | 69.87 | 38.94 | 0.8534 | 0.8102 | 1.182 | |
| SalamiTuning method=Fine-tuning, Tuning data=Dev set (Salami refs), Inference prompt=Salami prompt2026.01 | 47.48 | 69.87 | 38.94 | 0.8534 | 0.8102 | 1.182 | |
| TransLLaMA (Action)Tuning method=Fine-tuning, Tuning data=Dev set (Action-adapted refs), Inference prompt=-2026.01 | 40.31 | 60.43 | 45.87 | 0.8235 | 0.7842 | 1.173 | |
| TransLLaMA (Action)Tuning method=Fine-tuning, Tuning data=Dev set (Action-adapted refs), Inference prompt=–2026.01 | 40.31 | 60.43 | 45.87 | 0.8235 | 0.7842 | 1.173 | |
| TransLLaMA (Reference)Tuning method=Fine-tuning, Tuning data=Dev set (Reference), Inference prompt=-2026.01 | 39.52 | 59.61 | 47.09 | 0.8198 | 0.7839 | 1.193 | |
| TransLLaMA (Reference)Tuning method=Fine-tuning, Tuning data=Dev set (Reference), Inference prompt=–2026.01 | 39.52 | 59.61 | 47.09 | 0.8198 | 0.7839 | 1.193 | |
| TransLLaMA (Salami)Tuning method=Fine-tuning, Tuning data=Dev set (Salami refs), Inference prompt=-2026.01 | 39.12 | 59.67 | 46.68 | 0.8125 | 0.7798 | 1.187 | |
| TransLLaMA (Salami)Tuning method=Fine-tuning, Tuning data=Dev set (Salami refs), Inference prompt=–2026.01 | 39.12 | 59.67 | 46.68 | 0.8125 | 0.7798 | 1.187 | |
| Few-shot (Action)Tuning method=Prompt tuning, Tuning data=Dev set (Action-adapted refs), Inference prompt=Few-shot demonstrations2026.01 | 38.92 | 63.17 | 47.85 | 0.8241 | 0.7856 | 1.18 | |
| Few-shot (Action)Tuning method=Prompt tuning, Tuning data=Dev set (Action-adapted refs), Inference prompt=Few-shot demonstrations2026.01 | 38.92 | 63.17 | 47.85 | 0.8241 | 0.7856 | 1.18 | |
| Few-shot (Salami)Tuning method=Prompt tuning, Tuning data=Dev set (Salami refs), Inference prompt=Few-shot demonstrations2026.01 | 38.62 | 62.74 | 47.83 | 0.8245 | 0.7859 | 1.192 | |
| Few-shot (Salami)Tuning method=Prompt tuning, Tuning data=Dev set (Salami refs), Inference prompt=Few-shot demonstrations2026.01 | 38.62 | 62.74 | 47.83 | 0.8245 | 0.7859 | 1.192 | |
| Few-shot (Reference)Tuning method=Prompt tuning, Tuning data=Dev set (Reference), Inference prompt=Few-shot demonstrations2026.01 | 38.28 | 62.33 | 48 | 0.8301 | 0.7876 | 1.196 | |
| Few-shot (Reference)Tuning method=Prompt tuning, Tuning data=Dev set (Reference), Inference prompt=Few-shot demonstrations2026.01 | 38.28 | 62.33 | 48 | 0.8301 | 0.7876 | 1.196 |