List your company

Mathematical Reasoning on GSM8K (Accg, Accv, AUROC/APGR Analysis)

83AccgOLMo-7B† / Qwen-7B
22.6838.345469.66Mar 26, 2026

Evaluation results

MethodLinks
OLMo-7B† / Qwen-7BGenerator=OLMo-7B†, Verifier=Qwen-7B2026.038390776.353.874.880.89665.795.695
Llama-3-8B† / Qwen-7BGenerator=Llama-3-8B†, Verifier=Qwen-7B2026.0361902859.657.7594855.153.861.445.6
Mistral-7B† / Llama-3-8BGenerator=Mistral-7B†, Verifier=Llama-3-8B2026.0351611165.553.971.968.787.263.899.888.7
Llama-1B / Qwen-7BGenerator=Llama-1B, Verifier=Qwen-7B2026.0338905154.352.366.866.651.150.357.456.2
Llama-1B / Llama-3-8BGenerator=Llama-1B, Verifier=Llama-3-8B2026.0338612354.352.365.560.652.150.566.258.5
Qwen-0.5B / Qwen-7BGenerator=Qwen-0.5B, Verifier=Qwen-7B2026.0325906461.662.162.968.25252.152.853.9