ResearchBenchmarksArithmetic Reasoning on Synthetic ArithmeticFollow89.9AccuracyInfoRidge71.23276.078580.92585.7715Jul 7, 2025Evaluation ResultsMethodMethodLinksAccuracyInfoRidgeBackbone=LLaMA-3.1-8BBackbone=LLaMA-3.1-8B2025.0789.9InfoRidgeBackbone=Qwen-2.5-0.5BBackbone=Qwen-2.5-0.5B2025.0788.8InfoRidgeBackbone=GPT-2 MediumBackbone=GPT-2 Medium2025.0773.15InfoRidgeBackbone=GPT-2 SmallBackbone=GPT-2 Small2025.0771.95