Leuson Mario Pedro Da Silva, Jordan S. A. M. HI et Foutse Khomh
Ensemble de données (2025)
Abstract
Since the release of ChatGPT in November 2022, the landscape of developer Q&A platforms, particularly Stack Overflow, has undergone significant changes. The ability of large language models (LLMs) to generate immediate, human-like responses to technical questions has started discussions on their potential to replace traditional Q&A platforms. This dataset was collected as part of an empirical study analyzing Stack Overflow questions and evaluating responses generated by ChatGPT and LLaMA.
The dataset supports research aimed at:
Assessing the reliability of LLM-generated answers and their potential long-term impact on platforms like Stack Overflow.
Tracking the evolution of user engagement with Stack Overflow post-ChatGPT’s release. Comparing the performance of ChatGPT and LLaMA across different topics.
| Matériel d'accompagnement: | |
|---|---|
| Département: | Département de génie informatique et génie logiciel |
| URL de PolyPublie: | https://publications.polymtl.ca/64422/ |
| Source: | Zenodo |
| DOI: | 10.5281/zenodo.15086541 |
| Autres DOI associés à ce document: | 10.5281/zenodo.15086542 |
| URL officielle: | https://doi.org/10.5281/zenodo.15086541 |
| Date du dépôt: | 04 avr. 2025 16:54 |
| Dernière modification: | 30 janv. 2026 09:19 |
| Citer en APA 7: | Da Silva, L. M. P., S. A. M. HI, J., & Khomh, F. (2025). LLMs and Stack Overflow discussions: reliability, impact, and challenges [Ensemble de données]. Zenodo. https://doi.org/10.5281/zenodo.15086541 |
|---|---|
Statistiques
Dimensions
