Question Answering on Natural Questions (EN, ES, Gap)
18.8EN ScoreCorrected Evaluation
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Corrected EvaluationScale=8B, Shot=5-shot, Prompt=native, References=augmented2026.06 | 18.8 | 19.6 | 0.8 | |
| Standard EvaluationScale=8B, Shot=0-shot, References=original translated2026.06 | 15.2 | 15.8 | 0.58 | |
| Corrected EvaluationScale=3B, Shot=5-shot, Prompt=native, References=augmented2026.06 | 13.2 | 13.8 | 0.59 | |
| Standard EvaluationScale=3B, Shot=0-shot, References=original translated2026.06 | 10.9 | 9.7 | 1.22 | |
| Corrected EvaluationScale=1B, Shot=5-shot, Prompt=native, References=augmented2026.06 | 9.16 | 9.13 | 0.03 | |
| Standard EvaluationScale=1B, Shot=0-shot, References=original translated2026.06 | 6.99 | 4.77 | 2.22 |