Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Generative Performance on Ultrafeedback 61.1k (test)

69.8Win Rate

FA-DPO

50.45655.47860.565.522Nov 30, 2025
Updated 1mo ago

Evaluation Results

MethodLinks
2025.11
69.8
2025.11
69.8
2025.11
68.45
2025.11
67.8
2025.11
67.25
2025.11
67.05
2025.11
67
2025.11
66.9
2025.11
66.85
2025.11
66.85
2025.11
66.8
2025.11
66.65
2025.11
66.55
2025.11
66.05
2025.11
65.6
2025.11
65.55
2025.11
64.7
2025.11
64.4
2025.11
64.35
2025.11
62.6
2025.11
60.1
2025.11
57.85
2025.11
57.8
2025.11
56.65
2025.11
56.5
2025.11
56.35
2025.11
55.8
2025.11
54.9
2025.11
54.85
2025.11
51.2