Translation on English-Spanish
76.7AccuracyMTV
Evaluation Results
| Method | Links | |
|---|---|---|
| MTVBackbone=LLaMA-3-8B, shots=4, iterations=1002024.06 | 76.7 | |
| 400-shot ICLBackbone=LLaMA-3-8B, shots=400, Evaluation Protocol=In-context Learning2024.06 | 68.5 | |
| 10-shot ICLBackbone=LLaMA-3-8B, shots=10, Evaluation Protocol=In-context Learning2024.06 | 65.2 | |
| + LTVprompt=few-shot2025.02 | 53.9 | |
| + ICVprompt=few-shot2025.02 | 50.8 | |
| + SVprompt=few-shot2025.02 | 50.6 | |
| Transformerprompt=few-shot2025.02 | 50.4 | |
| + FVprompt=few-shot2025.02 | 49.4 | |
| Standard Promptingprompt=few-shot2025.02 | 36.7 | |
| + LTVprompt=zero-shot2025.02 | 20.7 | |
| + FVprompt=zero-shot2025.02 | 16.5 | |
| + ICVprompt=zero-shot2025.02 | 9.7 | |
| + SVprompt=zero-shot2025.02 | 9.4 | |
| Transformerprompt=zero-shot2025.02 | 9 | |
| Standard Promptingprompt=zero-shot2025.02 | 0 |