LLM Integration in Workbook Design for Teaching Coding Subjects

This work in progress paper explores the integration of Large Language Models (LLMs) into educational support tools, particularly in studying basics of Natural Language Processing. It investigates how generative LLMs can align with Bloom’s Taxonomy to support various learning levels, and how Prompt Engineering (PE) can improve communication with educational tools backed by LLMs. The approach involves planing integration LLMs into digital workbooks, employing techniques like Chain of Thoughts (CoT) and the Socratic method. These methods are utilized to provide tailored, interactive learning experiences that stimulate analytical thinking and deeper understanding of new concepts as well as providing instant response for coding exercises. Preliminary findings suggest that LLMs could significantly enhance the learning experience in coding education. The study suggests that PE combined with CoT and Socratic method could effectively addresses conversational problems with LLMs, enhancing learner engagement and understanding. This integration of LLMs facilitates a more dynamic form of learning, centered around dialogical exploration and practical experimentation, which aligns with the evolving objectives of lifelong learning. The ability to adapt and personalize knowledge acquisition to meet these ever-changing educational goals is a key advantage of this approach. The paper emphasizes the necessity for continuous research focused on the tailored integration of generative LLMs into educational tools. This involves employing PE techniques and leveraging a curated knowledge base to ensure that the LLMs’ responses are both relevant and contextually appropriate for educational purposes.