Shuai Wang

Other people with similar names: Shuai Wang , Shuai Wang , Shuai Wang


2024

pdf bib
OOP: Object-Oriented Programming Evaluation Benchmark for Large Language Models
Shuai Wang | Liang Ding | Li Shen | Yong Luo | Bo Du | Dacheng Tao
Findings of the Association for Computational Linguistics: ACL 2024

Advancing automated programming necessitates robust and comprehensive code generation benchmarks, yet current evaluation frameworks largely neglect object-oriented programming (OOP) in favour of functional programming (FP), e.g., HumanEval and MBPP. To address this, our study introduces a pioneering OOP-focused benchmark, featuring 431 Python programs that encompass essential OOP concepts and features like classes and encapsulation methods. We propose a novel evaluation metric, pass@o, tailored for OOP, enhancing traditional pass@k metric. Our evaluation of 23 leading large language models (LLMs), including both general and code-specialized models, reveals three key insights: 1) pass@o offers a more relevant and comprehensive assessment for OOP code generation; 2) Despite excelling in FP, code-specialized LLMs like WizardCoder lag in OOP compared to models like ChatGPT; 3) The poor performance of all advanced LLMs on our OOP benchmark highlights a critical need for improvements in this field. Our benchmark and scripts will be publicly released at GitHub.