Pietro Bernardelle

Room 622 - Building 78

The University of Queensland

Brisbane, Australia

I’m a second year Computer Science PhD Candidate at The University of Queensland working on bias in large language models (LLMs). Under the supervision of Professor Gianluca Demartini, Dr. Joel Mackenzie and Dr. Kevin Roitero, my current focus revolves around political bias in LLMs, exploring how persona-based conditioning can manipulate model outputs and developing bias mitigation strategies to enhance LLMs fairness.

My recent work include:

Developing a framework for reducing political bias in LLMs using DPO (ongoing).
Developing a framework for measuring LLMs perspective alignment effectiveness across downstream applications (ongoing).
Influencing political ideology in LLMs using synthetic personas (The Web Conference 2025).
Examining how data quality impacts LLMs alignment thorugh DPO (SIGIR-AP 2024).

selected publications

Ideology-Based LLMs for Content Moderation

Stefano Civelli, Pietro Bernardelle, Nardiena A. Pratama, and Gianluca Demartini

In ACM transactions on intelligent systems and technology, 2026

arXiv Bib

@inproceedings{civelli2025ideological,
  author = {Civelli, Stefano and Bernardelle, Pietro and Pratama, Nardiena A. and Demartini, Gianluca},
  title = {Ideology-Based LLMs for Content Moderation},
  booktitle = {ACM transactions on intelligent systems and technology},
  year = {2026}
}

Mapping and Influencing the Political Ideology of Large Language Models using Synthetic Personas

Pietro Bernardelle, Leon Fröhling, Stefano Civelli, Riccardo Lunardi, and 2 more authors

In Companion Proceedings of the ACM on Web Conference, 2025

DOI PDF arXiv Bib

@inproceedings{bernardelle2025mapping,
  title = {Mapping and Influencing the Political Ideology of Large Language Models using Synthetic Personas},
  author = {Bernardelle, Pietro and Fr\"{o}hling, Leon and Civelli, Stefano and Lunardi, Riccardo and Roitero, Kevin and Demartini, Gianluca},
  booktitle = {Companion Proceedings of the ACM on Web Conference},
  pages = {864--867},
  year = {2025}
}

The Impact of Persona-based Political Perspectives on Hateful Content Detection

Stefano Civelli, Pietro Bernardelle, and Gianluca Demartini

In Companion Proceedings of the ACM on Web Conference, 2025

DOI PDF arXiv Bib

@inproceedings{civelli2025impact,
  author = {Civelli, Stefano and Bernardelle, Pietro and Demartini, Gianluca},
  title = {The Impact of Persona-based Political Perspectives on Hateful Content Detection},
  booktitle = {Companion Proceedings of the ACM on Web Conference},
  pages = {1963--1968},
  year = {2025}
}