Question Answering on TriviaQA (EN, ES, Gap Metrics)
55.4Accuracy (EN)Corrected Evaluation
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Corrected EvaluationScale=8B, Shot=5-shot, Prompt=native, References=augmented2026.06 | 55.4 | 53.2 | 2.2 | |
| Standard EvaluationScale=8B, Shot=0-shot, References=original translated2026.06 | 53.5 | 47.4 | 6.12 | |
| Corrected EvaluationScale=3B, Shot=5-shot, Prompt=native, References=augmented2026.06 | 41.7 | 36.1 | 6.6 | |
| Standard EvaluationScale=3B, Shot=0-shot, References=original translated2026.06 | 40.6 | 31.7 | 8.87 | |
| Corrected EvaluationScale=1B, Shot=5-shot, Prompt=native, References=augmented2026.06 | 29.19 | 24.83 | 4.36 | |
| Standard EvaluationScale=1B, Shot=0-shot, References=original translated2026.06 | 27.68 | 21.56 | 6.12 |