Exploring the Robustness of Large Language Models for Solving Programming Problems
This addresses the reliability of LLMs for code generation, which is important for developers and educators, though it is incremental in exploring robustness.
The paper investigates whether large language models (LLMs) truly understand programming problem descriptions or rely on superficial cues from training data, finding that models like CodeGen and Codex are sensitive to modifications in prompts, while SOTA models like InstructGPT and ChatGPT show higher robustness.
Using large language models (LLMs) for source code has recently gained attention. LLMs, such as Transformer-based models like Codex and ChatGPT, have been shown to be highly capable of solving a wide range of programming problems. However, the extent to which LLMs understand problem descriptions and generate programs accordingly or just retrieve source code from the most relevant problem in training data based on superficial cues has not been discovered yet. To explore this research question, we conduct experiments to understand the robustness of several popular LLMs, CodeGen and GPT-3.5 series models, capable of tackling code generation tasks in introductory programming problems. Our experimental results show that CodeGen and Codex are sensitive to the superficial modifications of problem descriptions and significantly impact code generation performance. Furthermore, we observe that Codex relies on variable names, as randomized variables decrease the solved rate significantly. However, the state-of-the-art (SOTA) models, such as InstructGPT and ChatGPT, show higher robustness to superficial modifications and have an outstanding capability for solving programming problems. This highlights the fact that slight modifications to the prompts given to the LLMs can greatly affect code generation performance, and careful formatting of prompts is essential for high-quality code generation, while the SOTA models are becoming more robust to perturbations.