paper preprint

Chain-of-Thought Prompting Elicits Reasoning in Large Language Models

A preprint reported that providing intermediate reasoning exemplars in prompts can improve performance on multi-step reasoning tasks in sufficiently large language models.