List your company

Mathematical Word Problems on GSM8k (OOD)

72.9AccuracyR1 Distill -> GRPO
21.4234.78548.1561.515Dec 3, 2025

Evaluation results

MethodLinks
R1 Distill -> GRPOtraining=SFT + GRPO2025.1272.9
BOLT -> GRPOtraining=SFT + GRPO2025.1269.7
STaR -> GRPOtraining=SFT + GRPO2025.1268.6
SkillFactory -> GRPOtraining=SFT + GRPO2025.1268.2
RL-Onlytraining=RL2025.1267.7
R1 Distilltraining=SFT2025.1262.9
Qwen2.5 1.5B Instructtraining=None2025.1259.2
SkillFactorytraining=SFT2025.1259.1
STaRtraining=SFT2025.1231.1
BOLTtraining=SFT2025.1223.4