List your company

Mathematical Word Problem Solving on GSM8K

94.91Pass@1Learned
54.235664.795375.35585.9147Jun 12, 2025Jul 20, 2025Aug 28, 2025Oct 6, 2025Nov 13, 2025Dec 22, 2025Jan 30, 2026

Evaluation results

MethodLinks
LearnedBackbone Model=DS-R1-Qwen-7B, Cmax=3, Bmax=92162025.0694.91——
LearnedBackbone Model=DS-R1-Llama-8B, Cmax=3, Bmax=92162025.0693.78——
WaitBackbone Model=DS-R1-Qwen-7B, Cmax=3, Bmax=92162025.0691.74——
HmmBackbone Model=DS-R1-Qwen-7B, Cmax=3, Bmax=92162025.0691.23——
AlternativelyBackbone Model=DS-R1-Qwen-7B, Cmax=3, Bmax=92162025.0689.27——
CritiqueBackbone Model=DS-R1-Qwen-7B, Cmax=3, Bmax=92162025.0688.89——
Baseline (w/o BF)Backbone Model=DS-R1-Qwen-7B, Cmax=3, Bmax=92162025.0688.81——
HmmBackbone Model=DS-R1-Llama-8B, Cmax=3, Bmax=92162025.0685.22——
WaitBackbone Model=DS-R1-Llama-8B, Cmax=3, Bmax=92162025.0684.9——
LearnedBackbone Model=DS-R1-Qwen-1.5B, Cmax=3, Bmax=92162025.0683.68——
WaitBackbone Model=DS-R1-Qwen-1.5B, Cmax=3, Bmax=92162025.0680.14——
AlternativelyBackbone Model=DS-R1-Llama-8B, Cmax=3, Bmax=92162025.0679.82——
HmmBackbone Model=DS-R1-Qwen-1.5B, Cmax=3, Bmax=92162025.0679.47——
AlternativelyBackbone Model=DS-R1-Qwen-1.5B, Cmax=3, Bmax=92162025.0678.83——
Baseline (w/o BF)Backbone Model=DS-R1-Llama-8B, Cmax=3, Bmax=92162025.0678.43——
Baseline (w/o BF)Backbone Model=DS-R1-Qwen-1.5B, Cmax=3, Bmax=92162025.0678.08——
CritiqueBackbone Model=DS-R1-Llama-8B, Cmax=3, Bmax=92162025.0677.79——
CritiqueBackbone Model=DS-R1-Qwen-1.5B, Cmax=3, Bmax=92162025.0676.1——
TSLMevaluation_protocol=pass@12026.0161.6——
GRPOevaluation_protocol=pass@12026.0160.8——
Procedure Cloningevaluation_protocol=pass@12026.0155.9——
SCevaluation_protocol=pass@12026.0155.8——
AgentModel=Qwen3-4B, Setting=Agent2026.02——97.31
AgentModel=Qwen3-8B, Setting=Agent2026.02——98.95
AgentModel=Qwen3-14B, Setting=Agent2026.02——98.69
BCModel=Qwen3-4B, Setting=BC2026.02——97.01
BCModel=Qwen3-8B, Setting=BC2026.02——97.99
BCModel=Qwen3-14B, Setting=BC2026.02——98.71
IFPOModel=Qwen3-4B, Setting=IFPO2026.02——97.4
IFPOModel=Qwen3-8B, Setting=IFPO2026.02——97.62
IFPOModel=Qwen3-14B, Setting=IFPO2026.02——98.27
SafeChainModel=Qwen3-4B, Setting=SafeChain2026.02——95.65
SafeChainModel=Qwen3-8B, Setting=SafeChain2026.02——96.83
SafeChainModel=Qwen3-14B, Setting=SafeChain2026.02——99.03
STAR-1Model=Qwen3-4B, Setting=STAR-12026.02——96.55
STAR-1Model=Qwen3-8B, Setting=STAR-12026.02——98.42
STAR-1Model=Qwen3-14B, Setting=STAR-12026.02——98.37
STAR-1-mixModel=Qwen3-4B, Setting=STAR-1-mix2026.02——96.98
STAR-1-mixModel=Qwen3-8B, Setting=STAR-1-mix2026.02——98.56
STAR-1-mixModel=Qwen3-14B, Setting=STAR-1-mix2026.02——98.49
Tree-of-Thoughtevaluation_protocol=pass@1002026.01—62.3—