List your company

Knowledge on GPQA (GPQA, AVG)

36.36GPQA ScoreHT-MNPO
25.85628.58331.3134.037Sep 27, 2025

Evaluation results

MethodLinks
HT-MNPOReward Model=Skywork-Reward-V22025.0936.3671.8
TD-MNPO2025.0933.3371.08
SimPO2025.0932.3269.6
HT-MNPOReward Model=Athene-RM-8B2025.0930.8170.99
HT-MNPOReward Model=ArmoRM-Llama32025.0929.870.83
DPO2025.0929.2970.68
SFT Model2025.0928.2870.28
INPO2025.0927.7870.25
SPPO2025.0926.2670.19