Simultaneous Machine Translation on ACL60/60 En-Zh (eval)
100BLEUReference
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| ReferenceTuning method=–, Tuning data=–, Inference prompt=–2026.01 | 100 | 100 | 0 | 0.9582 | 0.8029 | 2.848 | |
| Step-wise (Drop / Cut)Tuning method=–, Tuning data=–, Inference prompt=Incremental drop / cut selection2026.01 | 62.84 | 44.06 | 104.8 | 0.8891 | 0.808 | 2.118 | |
| Step-wise (Actions)Tuning method=–, Tuning data=–, Inference prompt=Incremental action selection2026.01 | 62.44 | 46.8 | 126.2 | 0.9002 | 0.802 | 2.281 | |
| TransLLaMA (Action)Tuning method=Fine-tuning, Tuning data=Dev set (Action-adapted refs), Inference prompt=–2026.01 | 58.5 | 41.61 | 96.72 | 0.8826 | 0.8053 | 2.36 | |
| TransLLaMA (Salami)Tuning method=Fine-tuning, Tuning data=Dev set (Salami refs), Inference prompt=–2026.01 | 57.66 | 41.36 | 97.16 | 0.8768 | 0.795 | 2.376 | |
| TransLLaMA (Reference)Tuning method=Fine-tuning, Tuning data=Dev set (Reference), Inference prompt=–2026.01 | 57.66 | 41.27 | 97.16 | 0.8852 | 0.8 | 2.934 | |
| SalamiTuning method=Fine-tuning, Tuning data=Dev set (Salami refs), Inference prompt=Salami prompt2026.01 | 57.21 | 40.57 | 110.04 | 0.8705 | 0.7846 | 2.521 | |
| Few-shot (Action)Tuning method=Prompt tuning, Tuning data=Dev set (Action-adapted refs), Inference prompt=Few-shot demonstrations2026.01 | 55.8 | 50.31 | 105.9 | 0.8843 | 0.8079 | 2.395 | |
| Few-shot (Reference)Tuning method=Prompt tuning, Tuning data=Dev set (Reference), Inference prompt=Few-shot demonstrations2026.01 | 55.79 | 39.11 | 110.77 | 0.8856 | 0.804 | 2.945 | |
| Few-shot (Salami)Tuning method=Prompt tuning, Tuning data=Dev set (Salami refs), Inference prompt=Few-shot demonstrations2026.01 | 55.49 | 50.11 | 106.11 | 0.8779 | 0.7984 | 2.456 |