List your company

Math Reasoning on AMC '23 (Acc@16)

82.18Acc@16Ctrl-R
43.991253.905663.8273.7344Mar 2, 2026

Evaluation results

MethodLinks
Ctrl-RBackbone=Qwen3-8B-Base, beta=0.22026.0382.18
Reward ShapingBackbone=Qwen3-8B-Base2026.0381.87
NL GuidanceBackbone=Qwen3-8B-Base2026.0381.25
DAPOBackbone=Qwen3-8B-Base2026.0380.62
Ctrl-RBackbone=Qwen3-1.7B-Base, beta=0.22026.0349.69
Reward ShapingBackbone=Qwen3-1.7B-Base2026.0347.96
DAPOBackbone=Qwen3-1.7B-Base2026.0347.81
NL GuidanceBackbone=Qwen3-1.7B-Base2026.0347.65
Ctrl-RBackbone=Qwen3-1.7B-Base, beta=1, Importance Sampling (IS)=true2026.0346.71
Ctrl-RBackbone=Qwen3-1.7B-Base, beta=0, Importance Sampling (IS)=false2026.0345.46