List your company

Mathematical Reasoning on GSM8K (acc/mean@8)

90.03Accuracy (mean@8)DiPO
79.390882.152984.91587.6771Apr 15, 2026

Evaluation results

MethodLinks
DiPOBase Model=Llama3.1-8B-Instruct, Evaluation Protocol=maj@82026.0490.03
DAPOBase Model=Llama3.1-8B-Instruct, Evaluation Protocol=maj@82026.0489.89
DAPO w/ ELBase Model=Llama3.1-8B-Instruct, Evaluation Protocol=maj@82026.0489.66
CDEBase Model=Llama3.1-8B-Instruct, Evaluation Protocol=maj@82026.0489.01
Llama3.1-8B-InstructBase Model=Llama3.1-8B-Instruct, Evaluation Protocol=maj@82026.0479.8