Saved in:
Bibliographic Details
Main Author: Samant, Shaunak
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2512.22216
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917171782746112
author Samant, Shaunak
author_facet Samant, Shaunak
contents Automated program repair using neural models has shown promising results on benchmark datasets, yet practical deployment remains limited. In this study, we examine whether a small transformer model can meaningfully repair real-world Java bugs and whether syntactic correctness is a reliable proxy for semantic correctness. We fine-tune CodeT5-small (60.5M parameters) on 52,364 Java bug-fix pairs from CodeXGLUE and evaluate both token-level performance and syntactic validity using AST parsing. While the model converges cleanly and achieves high grammatical correctness, producing syntactically valid Java code in approximately ninety-four percent of cases, it fails to generate correct repairs under exact-match evaluation, achieving zero exact matches. In approximately eighty percent of cases, the model reproduces the buggy input verbatim.
format Preprint
id arxiv_https___arxiv_org_abs_2512_22216
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Syntax Is Not Enough: An Empirical Study of Small Transformer Models for Neural Code Repair
Samant, Shaunak
Software Engineering
Automated program repair using neural models has shown promising results on benchmark datasets, yet practical deployment remains limited. In this study, we examine whether a small transformer model can meaningfully repair real-world Java bugs and whether syntactic correctness is a reliable proxy for semantic correctness. We fine-tune CodeT5-small (60.5M parameters) on 52,364 Java bug-fix pairs from CodeXGLUE and evaluate both token-level performance and syntactic validity using AST parsing. While the model converges cleanly and achieves high grammatical correctness, producing syntactically valid Java code in approximately ninety-four percent of cases, it fails to generate correct repairs under exact-match evaluation, achieving zero exact matches. In approximately eighty percent of cases, the model reproduces the buggy input verbatim.
title Syntax Is Not Enough: An Empirical Study of Small Transformer Models for Neural Code Repair
topic Software Engineering
url https://arxiv.org/abs/2512.22216