Preference Classification on Anthropic HH Harmless (test)
71.7Accuracy
Few-Shot
Updated 4d ago
Evaluation Results
| Method | Links | ||
|---|---|---|---|
2026.02 | 71.7 | 0.082 | |
2026.02 | 71.5 | 0.06 | |
2026.02 | 71.5 | 0.061 | |
2026.02 | 71.2 | 0.043 | |
2026.02 | 71 | 0.089 | |
2026.02 | 70.9 | 0.108 | |
2026.02 | 70.7 | 0.077 | |
2026.02 | 70.1 | 0.107 | |
2026.02 | 69.5 | 0.04 | |
2026.02 | 69.1 | 0.064 | |
2026.02 | 68.9 | 0.064 | |
2026.02 | 68.5 | 0.072 | |
2026.02 | 67.6 | -0.01 | |
2026.02 | 67.4 | -0.024 | |
2026.02 | 63.8 | 0.012 | |
2025.11 | 58.4 | - | |
| 58 | - | ||
2025.11 | 58 | - | |
2025.11 | 57.1 | - | |
2025.11 | 56.4 | - | |
2025.11 | 55.4 | - | |
2025.11 | 51.2 | - |