Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Preference Alignment on UFB
Loading...
83.2
Win Rate
CW-DPO
54.184
61.717
69.25
76.783
Mar 5, 2026
Win Rate
Updated 1mo ago
Evaluation Results
Method
Method
Links
Win Rate
CW-DPO
Alignment Algorithm=DP...
2026.03
83.2
CW-rDPO
Alignment Algorithm=rD...
2026.03
79.8
WS-DPO
Alignment Algorithm=DP...
2026.03
79.5
Human
Alignment Algorithm=IP...
2026.03
77.1
Human
Alignment Algorithm=DP...
2026.03
76.8
CW-IPO
Alignment Algorithm=IP...
2026.03
76.6
WS-DPO
Alignment Algorithm=IP...
2026.03
75.4
WS-DPO
Alignment Algorithm=rD...
2026.03
74.3
Human
Alignment Algorithm=rD...
2026.03
70.9
WS-DPO
Alignment Algorithm=DP...
2026.03
64.7
CW-DPO
Alignment Algorithm=DP...
2026.03
64.3
CW-IPO
Alignment Algorithm=IP...
2026.03
63.9
Human
Alignment Algorithm=IP...
2026.03
62.1
WS-DPO
Alignment Algorithm=IP...
2026.03
60.2
Human
Alignment Algorithm=DP...
2026.03
59.8
CW-rDPO
Alignment Algorithm=rD...
2026.03
59.4
WS-DPO
Alignment Algorithm=rD...
2026.03
56.8
Human
Alignment Algorithm=rD...
2026.03
55.3
Feedback
Search any
task
Search any
task