PulseExploreJournal ClubDebatesTrendingResearchersJournals
Instagram
HomeExploreJournal ClubTrending
Synapse
⌘+K
Synapse
March 21, 20260 citationsOpen Access

Fine-Tuning CodeBERT-Java for Method-Level Bug Detection on Defects4J

View Full Paper
AAAnonymus AnonymusAAAnonymus Anonymus

Key Points

  • The aim is to fine-tune CodeBERT-Java for improving method-level bug detection using Defects4J data.
  • Developed a method-level bug detection dataset from buggy and fixed revisions of Defects4J projects.
  • Performed data extraction, preprocessing, and deduplication of changed Java methods.
  • Implemented dataset splitting and evaluated CodeBERT-Java models in zero-shot and fine-tuned scenarios.
  • Generated a dataset containing 2,574 labeled method instances for binary classification.
  • Evaluated the effectiveness of both zero-shot and fine-tuned models on the dataset.

Abstract

This package contains the scripts used to build and evaluate a method-level Java bug detection dataset derived from Defects4J. It supports data extraction from buggy and fixed revisions, preprocessing and deduplication of changed methods, dataset splitting, and evaluation of zero-shot and fine-tuned CodeBERT-Java models. The code was used on 16 Defects4J projects and the resulting dataset contains 2,574 labeled method instances for binary classification as buggy (1) or fixed (0).

Ask AI
Helpful
Bookmark
Share
View Full Paper

Cite This Study

Anonymus et al. (2026) studied this question.

synapsesocial.com/papers/69be37ce6e48c4981c677c2ehttps://doi.org/10.5281/zenodo.19087509
Ask AI
Helpful
Bookmark
Share
View Full Paper