List your company

Data Generation on GSM8K, Live Code Bench, and Dolly15k 53K labels

8.5Generation Time (hours)SelfJudge
4.0434.14564.2594.355Sep 26, 2025

Evaluation results

MethodLinks
SelfJudgeTarget/Draft Model=Llama-3.1-70B/8B, Hardware=4x H100 GPUs2025.098.5
AutoJudgeTarget/Draft Model=Llama-3.1-70B/8B, Hardware=4x H100 GPUs2025.09120