List your company

Mathematical Reasoning on GSM8K (Accuracy, Attack Robustness, Latency)

70Accuracy (Clean)AGLS
10.225.72541.2556.775May 26, 2025

Evaluation results

MethodLinks
AGLSModels=gemma2-27B2025.05701085.710.85
AGLSModels=gpt-4o-latest2025.0547.51568.422.21
AGLSModels=llama3.1-70B2025.0547.5589.471.72
AGLSModels=llama3.2-3B2025.0547.5589.471.38
AGLSModels=gpt-42025.0527.52.590.912.31
AGLSModels=qwen2.5-14B2025.0522.5577.782.49
AGLSModels=llama3.1-8B2025.0517.57.557.141.49
AGLSModels=qwen2.5-7B2025.05157.5501.91
AGLSModels=gemma2-9B2025.0512.57.5401.24