Explainable AI for Pre-Trained Code Models: What Do They Learn? When They Do Not Work?

TL;DR

This paper investigates what large language models for code learn and why they sometimes fail, using explainable AI methods to analyze their focus and limitations across various software engineering tasks.

Contribution

It provides an in-depth analysis of CodeBERT and GraphCodeBERT's learned representations and failure patterns using attention-based explainability techniques.

Findings

Models focus on specific source code token types.

Identified common patterns in model failures.

Recommendations for improving model robustness.

Abstract

In recent years, there has been a wide interest in designing deep neural network-based models that automate downstream software engineering tasks on source code, such as code document generation, code search, and program repair. Although the main objective of these studies is to improve the effectiveness of the downstream task, many studies only attempt to employ the next best neural network model, without a proper in-depth analysis of why a particular solution works or does not, on particular tasks or scenarios. In this paper, using an example eXplainable AI (XAI) method (attention mechanism), we study two recent large language models (LLMs) for code (CodeBERT and GraphCodeBERT) on a set of software engineering downstream tasks: code document generation (CDG), code refinement (CR), and code translation (CT). Through quantitative and qualitative studies, we identify what CodeBERT and GraphCodeBERT learn (put the highest attention on, in terms of source code token types), on these tasks. We also show some of the common patterns when the model does not work as expected (performs poorly even on easy problems) and suggest recommendations that may alleviate the observed challenges.