Pairwise Preference Ranking on UltraFeedback 2% holdout (test)
89.1Pairwise Acc (RM1-Honest)
BENCHALIGN
Updated 4d ago
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
2026.02 | 89.1 | 89.1 | 0.933 | 0.929 | |
2026.02 | 74.6 | 74.8 | 0.698 | 0.689 | |
2026.02 | 74.2 | 74.1 | 0.69 | 0.679 | |
2026.02 | 73.9 | 73.3 | 0.666 | 0.655 |