A Comparative Study of Prompting Techniques for LLMs in Educational Applications
Bourne Choi · 2025
With the advancement of Large Language Models (LLMs), they are being widely utilized in various fields. However, using LLMs directly in education can lead to the generation of hallucinated answers or responses with a structure unsuitable for learning. To address these issues, this study explores the most effective prompt methods for various categories of questions. Our experimental results show that the Chain-of-Thought (CoT) prompt was most effective for Math, Science, and History, while the Few-shot prompt was most effective for English.