CodeCoT: Tackling Code Syntax Errors in CoT Reasoning for Code Generation

TL;DR

CodeCoT enhances code generation by integrating chain-of-thought reasoning with self-examination and iterative refinement, significantly reducing syntax errors and improving execution success rates.

Contribution

This paper introduces CodeCoT, a novel method combining CoT reasoning with self-examination and test-based refinement to address syntax errors in code generation.

Findings

Increases pass@1 from 75.6% to 79.3% on HumanEval

Effectively reduces syntax errors during code execution

Demonstrates improved code correctness and reliability

Abstract

Chain-of-thought (CoT) has emerged as a groundbreaking tool in NLP, notably for its efficacy in complex reasoning tasks, such as mathematical proofs. However, its application in code generation faces a distinct challenge, i.e., although the code generated with CoT reasoning is logically correct, it faces the problem of syntax error (e.g., invalid syntax error report) during code execution, which causes the CoT result's pass@1 in HumanEval even lower than the zero-shot result. In this paper, we present Code Chain-of-Thought (CodeCoT) that integrates CoT with a self-examination process for code generation. CodeCoT begins with the LLMs using CoT for initial code development to ensure the generated code follows the correct logic flow. Then, CodeCoT will generate test cases to validate whether the code has syntax errors during the execution. CodeCoT then employs a self-examination phase, in which the generated code is executed against these test cases in the local environment. If the local environment raises error information (e.g., invalid syntax error), CodeCoT will iteratively refine the code based on the feedback information. Within this loop, CodeCoT can make sure their generated codes not only follow the logic flow of the code description, but the syntax error will also be addressed with the self-examination process. Our evaluation results reveal that CodeCoT improves the effectiveness of code generation. For example, CodeCoT increases pass@1 from 75.6% to 79.3% for the HumanEval dataset.