零样本通识推理的提示工程和校准（英文版）

英文标题：Prompt Engineering and Calibration for Zero-Shot Commonsense Reasoning中文摘要：在小型语言模型上研究及评估大型语言模型的理解任务，包括多项选择常识推理，结果显示各种策略对不同的模型有利，但联合效果大多为负。英文摘要：Prompt engineering and calibration make large language models excel atreasoning tasks, including multiple choice commonsense reasoning. From apractical perspective, we investigate and evaluate these strategies on smallerlanguage models. Through experiments on five commonsense reasoning benchmarks,we find that each strategy favors certain models, but