Automated Bug Triaging using Instruction-Tuned Large Language Models

TL;DR

This paper introduces a lightweight, instruction-tuned large language model framework for bug triaging that improves assignment quality and shows promise for real-world applications, reducing reliance on complex feature engineering.

Contribution

The paper presents a novel framework using instruction-tuned LLMs with LoRA adapters and candidate-constrained decoding for bug triaging, demonstrating effectiveness on real datasets.

Findings

Achieves high shortlist quality (Hit@10 up to 0.753)

Accuracy improves significantly on recent snapshots

Offers a practical alternative to traditional feature engineering methods

Abstract

Bug triaging, the task of assigning new issues to developers, is often slow and inconsistent in large projects. We present a lightweight framework that instruction-tuned large language model (LLM) with LoRA adapters and uses candidate-constrained decoding to ensure valid assignments. Tested on EclipseJDT and Mozilla datasets, the model achieves strong shortlist quality (Hit at 10 up to 0.753) despite modest exact Top-1 accuracy. On recent snapshots, accuracy rises sharply, showing the framework's potential for real-world, human-in-the-loop triaging. Our results suggest that instruction-tuned LLMs offer a practical alternative to costly feature engineering and graph-based methods.