cff-version: 1.2.0
message: "If you use AIFaultBench, please cite it as below."
type: dataset
title: "AIFaultBench: A Reproducible Benchmark of Real-World AI Software Faults"
abstract: >-
  AIFaultBench is a benchmark of 770 real-world AI software faults collected from
  105 open-source repositories across 76 organizations, spanning traditional
  machine learning, deep learning, large language model infrastructure,
  reinforcement learning, agentic AI systems, and AI tooling. Each fault ships
  with the original GitHub issue report, a minimal reproduction script, a
  dependency specification, codebase reconstruction and environment setup
  scripts, reproduction logs, structured metadata, and a reproduction
  trajectory. 652 of the 770 faults (85%) are verified reproducible; the
  remainder document the reasons preventing reproduction.
authors:
  - family-names: Shah
    given-names: "Mehil B."
    email: shahmehil@dal.ca
    affiliation: "Dalhousie University"
  - family-names: Rahman
    given-names: "Mohammad Masudur"
    affiliation: "Dalhousie University"
  - family-names: Khomh
    given-names: Foutse
    affiliation: "Polytechnique Montréal"
doi: 10.5281/zenodo.21763133
identifiers:
  - type: doi
    value: 10.5281/zenodo.21763133
    description: "Version DOI for v1.1."
  - type: doi
    value: 10.5281/zenodo.21763132
    description: "Concept DOI — always resolves to the latest version."
url: "https://zenodo.org/records/21763203"
repository-code: "https://github.com/mehilshah/AIFaultBench"
repository-artifact: "https://huggingface.co/datasets/mehilshah/AIFaultBench"
license: CC-BY-4.0
version: v1.1
date-released: 2026-08-02
keywords:
  - benchmark
  - software engineering
  - bug reproduction
  - fault localization
  - automated program repair
  - debugging
  - machine learning
  - deep learning
  - large language models
  - agentic AI
  - reinforcement learning
  - mining software repositories