List your company

Mathematical Reasoning on GSM8K (test) (Accuracy, Tokens/sec)

65.4Throughput (Tokens/sec)SelfJudge-F
52.97256.198559.42562.6515Sep 26, 2025

Evaluation results

MethodLinks
SelfJudge-FModel (Target/Draft)=Llama-3.1-70B/8B, Hardware=4 A100 GPUs, Tensor Parallelism=4-way, Draft length (gamma)=202025.0965.494
SelfJudge-RModel (Target/Draft)=Llama-3.1-70B/8B, Hardware=4 A100 GPUs, Tensor Parallelism=4-way, Draft length (gamma)=202025.0959.5793.8
SDModel (Target/Draft)=Llama-3.1-70B/8B, Hardware=4 A100 GPUs, Tensor Parallelism=4-way, Draft length (gamma)=202025.0953.4593