JavaLLM: A Fine-Tuned LLM for Java Programming Education
Jingying Zhang, Kang Liu · 2024
The integration of Large Language Models (LLMs) into education marks a significant advancement toward personalized and adaptive learning environments, particularly in programming education. Addressing the limitations of existing LLMs in specialized domains like Java programming, this paper introduces JavaLLM — a model specifically tailored for Java programming education. Built upon a robust codeLLM and fine-tuned using extensive, high-quality Java-focused datasets, JavaLLM demonstrates superior performance in code generation and Java-specific question answering. Through rigorous evaluation and iterative refinement, JavaLLM facilitates a transformative classroom experience, enhancing the quality of teaching and enabling a personalized learning journey for students in Java programming courses. This innovation paves the way for smarter, more tailored educational approaches, leveraging AI’s generative capabilities to meet the evolving demands of modern education.