List your company

Reward model verification on HH-RLHF

47.3Win RateAISP
24.10830.12936.1542.171Oct 30, 2025

Evaluation results

MethodLinks
AISPBase Model=Gemma3 4B, Reward Model=Eurus2025.1047.3———
AISPBase Model=Llama3 8B, Reward Model=Eurus2025.1046.3———
AISPBase Model=Gemma3 4B, Reward Model=UltraRM2025.1044———
BoN (top-p)Base Model=Llama3 8B, Reward Model=Eurus2025.1043.7———
BoN (top-p)Base Model=Gemma3 4B, Reward Model=Eurus2025.1041.7———
AISPBase Model=Llama3 8B, Reward Model=UltraRM2025.1041———
BoN (top-p)Base Model=Llama3 8B, Reward Model=UltraRM2025.1041———
BoN (top-p)Base Model=Gemma3 4B, Reward Model=UltraRM2025.1036.7———
BoN (top-p)Base Model=Vicuna 7B, Reward Model=Eurus2025.1033———
AISPBase Model=Vicuna 7B, Reward Model=Eurus2025.1032.3———
AISPBase Model=Vicuna 7B, Reward Model=UltraRM2025.1025———
BoN (top-p)Base Model=Vicuna 7B, Reward Model=UltraRM2025.1025———
BT RM2025.05—0.74380.73230.7381
FC RM2025.05—0.7430.73620.7396
FR RM2025.05—0.73490.73620.7355