List your company

Mathematical Reasoning on GSM8K (test) (pass@1 accuracy)

78.3Pass@1 AccuracyLearnAlign
54.79660.8986773.102Jun 13, 2025

Evaluation results

MethodLinks
LearnAlignSelected Data Size=2,0002025.0678.3
LearnAlignSelected Data Size=1,0002025.0677.5
Qwen2.5-1.5B-Instruct-FULL2025.0677
LIMRSelected Data Size=2,0002025.0676.7
LearnAlignSelected Data Size=5002025.0676.4
Token LengthSelected Data Size=1,0002025.0676.2
LIMRSelected Data Size=5002025.0676.2
LIMRSelected Data Size=1,0002025.0676.1
IFDSelected Data Size=5002025.0676
PPL-TopSelected Data Size=5002025.0675.8
SelectITSelected Data Size=5002025.0675.7
Random SamplingSelected Data Size=1,0002025.0675.6
IFDSelected Data Size=1,0002025.0675.6
Token LengthSelected Data Size=2,0002025.0675.6
SelectITSelected Data Size=1,0002025.0675.6
Random SamplingSelected Data Size=2,0002025.0675.5
SelectITSelected Data Size=2,0002025.0675.5
IFDSelected Data Size=2,0002025.0675.4
PPL-TopSelected Data Size=2,0002025.0675.2
Random SamplingSelected Data Size=5002025.0675.1
PPL-MiddleSelected Data Size=1,0002025.0675
LearnAlignSelected Data Size=1002025.0674.8
PPL-MiddleSelected Data Size=5002025.0674.7
PPL-TopSelected Data Size=1,0002025.0674.6
Token LengthSelected Data Size=5002025.0674.4
PPL-MiddleSelected Data Size=2,0002025.0674.2
LIMRSelected Data Size=1002025.0674.2
Random SamplingSelected Data Size=1002025.0673.1
PPL-MiddleSelected Data Size=1002025.0672.8
SelectITSelected Data Size=1002025.0672.8
PPL-TopSelected Data Size=1002025.0672.5
Token LengthSelected Data Size=1002025.0672.3
IFDSelected Data Size=1002025.0672
Qwen2.5-1.5B-Instruct2025.0655.7