California Driver's Handbook Context

该数据集包含了从加利福尼亚州驾照知识测试中选取的50个样本问题，用于评估在没有和有加利福尼亚州驾驶手册上下文信息的情况下，GPT-3模型的表现性能。数据集还包括了根据提供给模型的不同数量和格式的上下文信息所表现出的性能变化。规模上，该数据集包含了50个问题，任务是对GPT-3模型在驾照知识测试问题上的测试性能进行评估。

This dataset includes 50 sample questions selected from the California Driver's License Knowledge Test, which is utilized to assess the performance of the GPT-3 model both in the absence and presence of contextual information from the California Driver's Handbook. It also records the variations in model performance corresponding to different quantities and formats of contextual information provided to the model. In terms of scale, this dataset contains 50 questions, and its core task is to evaluate the test performance of the GPT-3 model on questions from the driver's license knowledge test.