Core Concepts
Large Language Models (LLMs) struggle with object-oriented programming, emphasizing the importance of code quality and supervision in educational settings.
Abstract
Large Language Models (LLMs) like GPT-3.5, GPT-4, and Bard are evaluated for handling Object-Oriented Programming (OOP) assignments.
While these models can provide solutions, they often overlook best practices of OOP.
GPT-4 outperformed GPT-3.5 and Bard, but all models faced challenges in adhering to OOP principles.
Recommendations include focusing on code quality, using LLMs in classes with supervision, and adopting project-based learning.
The study highlights the need for curated sources and understanding the limitations of LLMs in educational contexts.
Stats
대규모 언어 모델 (LLMs)는 객체 지향 프로그래밍 (OOP) 과제를 처리하는 데 어려움을 겪습니다.
GPT-4는 GPT-3.5와 Bard보다 우수한 성과를 보였지만, 모든 모델이 OOP 원칙을 준수하는 데 어려움을 겪었습니다.
권고 사항은 코드 품질에 중점을 두고 감독하며 수업에서 LLM 사용 및 프로젝트 기반 학습을 채택하는 것을 포함합니다.
Quotes
"GPT-4 stood out as the most proficient, followed by GPT-3.5, with Bard trailing last."
"We advocate for a renewed emphasis on code quality when employing these models and explore the potential of pairing LLMs with AATs in pedagogical settings."
"These experiments show that these 3 LLMs were able to partially solve the ‘IT Company’ assignment, although with some errors."