Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Alignment on HH (IDN 20%)
Loading...
78.2
Win Rate
α-UDPO
55.944
61.722
67.5
73.278
Jul 3, 2026
Win Rate
Updated 18d ago
Evaluation Results
Method
Method
Links
Win Rate
α-UDPO
Base Model=Qwen-3-1.7B
2026.07
78.2
UDPO
Base Model=Qwen-3-1.7B
2026.07
77.9
rDPO
Base Model=Qwen-3-1.7B
2026.07
73.3
DPO
Base Model=Qwen-3-1.7B
2026.07
68.7
α-URM
Base Model=Qwen-3-1.7B
2026.07
60.2
URM
Base Model=Qwen-3-1.7B
2026.07
59.5
rRM
Base Model=Qwen-3-1.7B
2026.07
58
RM
Base Model=Qwen-3-1.7B
2026.07
56.8
Feedback
Search any
task
Search any
task