List your company

Mathematical Reasoning on GSM8K (GSM8K Score)

90GSM8K ScoreGRPO
77.5280.768487.24May 26, 2025

Evaluation results

MethodLinks
GRPOBackbone=LLaMA-3.1-8B2025.0590
TI-DPOBackbone=LLaMA-3.1-8B2025.0589
TPOBackbone=LLaMA-3.1-8B2025.0588
CPOBackbone=LLaMA-3.1-8B2025.0586.5
KTOBackbone=LLaMA-3.1-8B2025.0586
DPOBackbone=LLaMA-3.1-8B2025.0585
SFTBackbone=LLaMA-3.1-8B2025.0584
TDPOBackbone=LLaMA-3.1-8B2025.0583
IPOBackbone=LLaMA-3.1-8B2025.0580
SIMPOBackbone=LLaMA-3.1-8B2025.0578