List your company

Mathematical Reasoning on GSM8K (test) (Accuracy, Speed-up)

55.9AccuracyVinePPO
50.90852.20453.554.796Feb 17, 2025

Evaluation results

MethodLinks
VinePPOLILO Integration=true2025.0255.93.3
PPOLILO Integration=true2025.0253.21.9
VinePPOLILO Integration=false2025.0253.2—
PPOLILO Integration=false2025.0251.1—