Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Test-time Alignment on HH-RLHF
Loading...
86
Win Rate vs Llama
CAST
85.272
85.461
85.65
85.839
Apr 26, 2026
Win Rate vs Llama
Win Rate vs DeepSeek
Win Rate vs GPT
Updated 2mo ago
Evaluation Results
Method
Method
Links
Win Rate vs Llama
Win Rate vs DeepSeek
Win Rate vs GPT
CAST
Base Model=Hermes3-8B,...
2026.04
86
79.9
74.7
Pref-CTRL
Base Model=Hermes3-8B,...
2026.04
85.7
84.3
73.6
Best-of-N
Base Model=Hermes3-8B,...
2026.04
85.3
78.9
72.1
Feedback
Search any
task
Search any
task