Safety and Informativeness Evaluation on OrBench Hard
75Deception Rate (Safe)
Mistral MoE-XL
Updated 1d ago
Evaluation Results
| Method | Links | ||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2026.05 | 75 | 72.7 | 90 | 63.3 | 80 | 66.9 | 90 | 63.9 | 90 | 70.6 | 90 | 76.7 | 85 | 73 | 90 | 70 | 95 | 72.1 | 96 | 72.2 | 87.1 | 70.1 | |
2026.05 | 65 | 65 | 71 | 60 | 70 | 67.5 | 85 | 67 | 54 | 62.4 | 80 | 68.4 | 90 | 68 | 90 | 65 | 80 | 65 | 76 | 70 | 73.4 | 65.3 | |
2026.05 | 48.1 | 50.8 | 54.3 | 44.2 | 58.8 | 48.7 | 73.3 | 55.2 | 51.9 | 46.6 | 61 | 49.1 | 58.8 | 51.4 | 91.8 | 45.4 | 70.6 | 51 | 36.4 | 63.3 | 59.41 | 51.4 | |
2026.05 | 41.3 | 41.9 | 56.2 | 41.8 | 49.4 | 40.9 | 64.1 | 44.3 | 53.2 | 40 | 42.4 | 46.7 | 62.7 | 41.9 | 89.3 | 39.8 | 55.4 | 38.2 | 44.4 | 37.4 | 55.84 | 41.3 |