List your company

Mathematical Reasoning on AMC 2023 (test) (pass@1)

65.4Pass@1 RateGRPO + coupled rhythm credit
54.68857.46960.2563.031Oct 15, 2025

Evaluation results

MethodLinks
GRPO + coupled rhythm creditBackbone=Qwen3-8B-Base, Context Length=1K2025.1065.4
GRPO + global-anchor creditBackbone=Qwen3-8B-Base, Context Length=1K2025.1063.1
GRPO + local-chunk creditBackbone=Qwen3-8B-Base, Context Length=1K2025.1062.9
GRPO + global-anchor creditBackbone=Qwen3-4B-Base, Context Length=8K2025.1061.1
GRPO + high-entropy creditBackbone=Qwen3-8B-Base, Context Length=1K2025.1060.4
GRPO + random creditBackbone=Qwen3-8B-Base, Context Length=1K2025.1060.1
GRPO + coupled rhythm creditBackbone=Qwen3-4B-Base, Context Length=8K2025.1059.3
GRPO + local-chunk creditBackbone=Qwen3-4B-Base, Context Length=8K2025.1059.2
GRPOBackbone=Qwen3-8B-Base, Context Length=1K2025.1059.1
GRPO + local-chunk creditBackbone=Qwen3-4B-Base, Context Length=1K2025.1058.4
GRPO + high-entropy creditBackbone=Qwen3-4B-Base, Context Length=8K2025.1057.8
GRPO + global-anchor creditBackbone=Qwen3-4B-Base, Context Length=1K2025.1057.6
GRPOBackbone=Qwen3-4B-Base, Context Length=8K2025.1057.6
GRPO + coupled rhythm creditBackbone=Qwen3-4B-Base, Context Length=1K2025.1057.4
GRPO + random creditBackbone=Qwen3-4B-Base, Context Length=8K2025.1057.1
GRPO + high-entropy creditBackbone=Qwen3-4B-Base, Context Length=1K2025.1055.5
GRPO + random creditBackbone=Qwen3-4B-Base, Context Length=1K2025.1055.2
GRPOBackbone=Qwen3-4B-Base, Context Length=1K2025.1055.1