ResearchBenchmarksMathematical Reasoning on GSM8K STREETFollow94Answer AccuracyGPT47.05629.62852.274.772Jan 24, 2024Evaluation ResultsMethodMethodLinksAnswer AccuracyReasoning Graph AccuracyGPT4model_version=gpt-4-11...model_version=gpt-4-1106-preview, prompting_strategy=best of CoT, ToT, and ReAct2024.019410SEER2024.0121.413.4STREET2024.0110.40.7