Nafiseh Nikeghbal

Chair of Human-Centered Computing

School of Computation, Information and Technology

Munich, Germany

I am a PhD candidate in Computer Science at the Technical University of Munich (TUM), where I am advised by Prof. Dr. Jana Diesner at the Chair of Human-Centered Computing. My research sits at the intersection of AI safety, societal bias in AI, natural language processing, and computational social science.

I build methods to understand and mitigate the harms of large language models — probing when and how models reveal hidden societal biases, evaluating their behavior across languages and modalities, and studying their reliability under adversarial and interactive conditions. My recent work includes CoBia (EMNLP 2025), which uses constructed conversations to surface concealed biases in LLMs, and MEXA (ACL Findings 2025), a method for evaluating the multilingual capabilities of English-centric LLMs.

Previously, I received my M.Sc. in Computer Engineering from Sharif University of Technology and my B.Sc. in Computer Engineering from Ferdowsi University of Mashhad. Feel free to reach out at nafiseh.nikeghbal@tum.de.

news

Aug 20, 2026	Our paper Who Flips? Self- and Cross-Model Counterarguments Reveal Answer Instability in LLMs was accepted to Findings of EMNLP 2026.
Jun 01, 2026	StylisticBias was accepted to the AI for Good (AI4GOOD) Workshop at ICML 2026.
Aug 21, 2025	CoBia was accepted to EMNLP 2025 (main conference) as an oral presentation. See the paper and code.
May 15, 2025	MEXA was accepted to the Findings of ACL 2025. Check out the paper and leaderboard.
Oct 01, 2024	I started my PhD in Computer Science at the Technical University of Munich, advised by Prof. Dr. Jana Diesner at the Chair of Human-Centered Computing.

selected publications

EMNLP
Who Flips? Self- and Cross-Model Counterarguments Reveal Answer Instability in LLMs

Nafiseh Nikeghbal, Amir Hossein Kargaran, Shaghayegh Kolli, and 1 more author

In Findings of the Association for Computational Linguistics: EMNLP 2026, Nov 2026

Abs Bib

We study how robust large language models are to counterarguments, showing that self- and cross-model counterarguments can flip model answers and reveal underlying answer instability.
@inproceedings{nikeghbal2026whoflips, title = {Who Flips? Self- and Cross-Model Counterarguments Reveal Answer Instability in LLMs}, author = {Nikeghbal, Nafiseh and Kargaran, Amir Hossein and Kolli, Shaghayegh and Diesner, Jana}, booktitle = {Findings of the Association for Computational Linguistics: EMNLP 2026}, year = {2026}, month = nov, }

ACL

MEXA: Multilingual Evaluation of English-Centric LLMs via Cross-Lingual Alignment

Amir Hossein Kargaran, Ali Modarressi, Nafiseh Nikeghbal, and 3 more authors

In Findings of the Association for Computational Linguistics: ACL 2025, Jul 2025

Abs arXiv Bib HTML Code

@inproceedings{kargaran2025mexa,
  title = {MEXA: Multilingual Evaluation of English-Centric LLMs via Cross-Lingual Alignment},
  author = {Kargaran, Amir Hossein and Modarressi, Ali and Nikeghbal, Nafiseh and Diesner, Jana and Yvon, Fran{\c{c}}ois and Sch{\"u}tze, Hinrich},
  booktitle = {Findings of the Association for Computational Linguistics: ACL 2025},
  pages = {27001--27023},
  year = {2025},
  month = jul,
  publisher = {Association for Computational Linguistics},
  address = {Vienna, Austria},
}

EMNLP
CoBia: Constructed Conversations Can Trigger Otherwise Concealed Societal Biases in LLMs

Nafiseh Nikeghbal, Amir Hossein Kargaran, and Jana Diesner

In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing (EMNLP), Nov 2025

Abs arXiv Bib HTML Code

CoBia is a suite of lightweight adversarial attacks that construct conversations in which a model utters a biased claim about a social group, then tests whether the model can recover and reject biased follow-ups. Across 11 open and proprietary LLMs and six socio-demographic categories, constructed conversations reliably reveal biases that standard safety checks miss.
@inproceedings{nikeghbal2025cobia, title = {CoBia: Constructed Conversations Can Trigger Otherwise Concealed Societal Biases in LLMs}, author = {Nikeghbal, Nafiseh and Kargaran, Amir Hossein and Diesner, Jana}, booktitle = {Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing (EMNLP)}, year = {2025}, month = nov, publisher = {Association for Computational Linguistics}, }