List your company

Test-time Alignment on HH-RLHF

86Win Rate vs LlamaCAST
85.27285.46185.6585.839Apr 26, 2026

Evaluation results

MethodLinks
CASTBase Model=Hermes3-8B, Alignment Stage=Test-time2026.048679.974.7
Pref-CTRLBase Model=Hermes3-8B, Alignment Stage=Test-time2026.0485.784.373.6
Best-of-NBase Model=Hermes3-8B, Alignment Stage=Test-time2026.0485.378.972.1