Pairwise Ranking
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
77.24Pairwise Preference Accuracy
13
Mar 5, 2026
4,950Average Human Annotation Count
12
Mar 24, 2026
62Accuracy (w/ Tie)
11
Feb 26, 2026
99.7NDCG
4
Feb 26, 2026
0.999NDCG
4
Feb 26, 2026
83.1Insult Pairwise Score
3
Mar 4, 2026
100Accuracy
3
Feb 26, 2026
0.991NDCG
2
Feb 26, 2026
98.3NDCG
2
Feb 26, 2026