Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

Discriminative Performance on Ultrafeedback 61.1k (test)

73.05Accuracy

FA-DPO

46.654853.507460.3667.2126Nov 30, 2025
Updated 4d ago

Evaluation Results

MethodLinks
2025.11
73.05
2025.11
70.97
2025.11
70.77
2025.11
70.75
2025.11
70.13
2025.11
69.77
2025.11
68.22
2025.11
67.2
2025.11
67.2
2025.11
65.9
2025.11
64.93
2025.11
63.73
2025.11
63.64
2025.11
62.57
2025.11
61.77
2025.11
61.7
2025.11
59
2025.11
58.93
2025.11
58.63
2025.11
57.81
2025.11
56.87
2025.11
56.67
2025.11
55.53
2025.11
55.43
2025.11
55.23
2025.11
54.07
2025.11
53.93
2025.11
51.87
2025.11
51.77
2025.11
47.67