BugPilot: Complex Bug Generation for Efficient Learning of SWE Skills

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Sonwane, Atharv, White, Isadora, Lee, Hyunji, Pereira, Matheus, Caccia, Lucas, Kim, Minseon, Shi, Zhengyan, Singh, Chinmay, Sordoni, Alessandro, Côté, Marc-Alexandre, Yuan, Xingdi
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908618143563776
author Sonwane, Atharv
White, Isadora
Lee, Hyunji
Pereira, Matheus
Caccia, Lucas
Kim, Minseon
Shi, Zhengyan
Singh, Chinmay
Sordoni, Alessandro
Côté, Marc-Alexandre
Yuan, Xingdi
author_facet Sonwane, Atharv
White, Isadora
Lee, Hyunji
Pereira, Matheus
Caccia, Lucas
Kim, Minseon
Shi, Zhengyan
Singh, Chinmay
Sordoni, Alessandro
Côté, Marc-Alexandre
Yuan, Xingdi
contents High quality bugs are key to training the next generation of language model based software engineering (SWE) agents. We introduce a novel method for synthetic generation of difficult and diverse bugs. Our method instructs SWE Agents to introduce a feature into the codebase whereby they may unintentionally break tests, resulting in bugs. Prior approaches often induce an out-of-distribution effect by generating bugs intentionally (e.g. by introducing local perturbation to existing code), which does not reflect realistic development processes. We perform qualitative analysis to demonstrate that our approach for generating bugs more closely reflects the patterns found in human-authored edits. Through extensive experiments, we demonstrate that our bugs provide more efficient training data for supervised fine-tuning, outperforming other bug datasets by 2% with half the training data (1.2k vs. 3k bugs). We train on our newly generated bugs in addition to existing bug datasets to get FrogBoss a state-of-the-art 32B parameter model on SWE-bench Verified with a pass@1 of 54.6% and FrogMini a state-of-the-art 14B model on SWE-bench Verified with a pass@1 of 45.3% on SWE-bench Verified averaged over three seeds.
format Preprint
id arxiv_https___arxiv_org_abs_2510_19898
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle BugPilot: Complex Bug Generation for Efficient Learning of SWE Skills
Sonwane, Atharv
White, Isadora
Lee, Hyunji
Pereira, Matheus
Caccia, Lucas
Kim, Minseon
Shi, Zhengyan
Singh, Chinmay
Sordoni, Alessandro
Côté, Marc-Alexandre
Yuan, Xingdi
Software Engineering
Artificial Intelligence
Computation and Language
High quality bugs are key to training the next generation of language model based software engineering (SWE) agents. We introduce a novel method for synthetic generation of difficult and diverse bugs. Our method instructs SWE Agents to introduce a feature into the codebase whereby they may unintentionally break tests, resulting in bugs. Prior approaches often induce an out-of-distribution effect by generating bugs intentionally (e.g. by introducing local perturbation to existing code), which does not reflect realistic development processes. We perform qualitative analysis to demonstrate that our approach for generating bugs more closely reflects the patterns found in human-authored edits. Through extensive experiments, we demonstrate that our bugs provide more efficient training data for supervised fine-tuning, outperforming other bug datasets by 2% with half the training data (1.2k vs. 3k bugs). We train on our newly generated bugs in addition to existing bug datasets to get FrogBoss a state-of-the-art 32B parameter model on SWE-bench Verified with a pass@1 of 54.6% and FrogMini a state-of-the-art 14B model on SWE-bench Verified with a pass@1 of 45.3% on SWE-bench Verified averaged over three seeds.
title BugPilot: Complex Bug Generation for Efficient Learning of SWE Skills
topic Software Engineering
Artificial Intelligence
Computation and Language
url https://arxiv.org/abs/2510.19898