Hello! 🐈‍⬛

University of British Columbia · alexf01@cs.ubc.ca

I am a master's student at the University of British Columbia advised by Peter West, where I work on analysis and understanding of language models.

During my undergrad, I was a visiting researcher at Johns Hopkins University advised by Najim Dehak and Thomas Thebaud. While visiting JHU, I had the opportunity to work on very cool speech stuff; such as backdoor attacks on speech language models. I completed my B.Eng at ÉTS in Montréal (where I am from!), advised and mentored by Patrick Cardinal.

My research interests are numerous and everchanging. At the moment, I think about how knowledge is learned, transformed and shaped through the training stages (from pretraining to post-training). More broadly, I also think about LLMs under a natural science lens. For example, are LLMs capable of creativity? Or should the term creativity be reserved to sentient beings?

I love to talk about anything (really); please do not hesitate to reach out!

Alexandrine Fortier

Click for a glimpse of my life!

News

2026.10
Will be attending CoLM 2026 in San Francisco!
2026.09
I will be giving an oral to present our work on backdoor attacks against speech LMs at Interspeech 2026 in Sydney 🦘
2026.09
TAing for Peter's course Open Qs this term!
2026.08
Our paper was accepted to EMNLP 2026! Our work traces back output homogeneity to base models, we ask ourselves: Is convergence inevitable in LMs 🤔
2025.12
Presented my attack work at ASRU 2025 in Honolulu!
2025.09
I started my master's at UBC. Vancouver is lovely 🌄

Research

Is Convergence Inevitable? Tracing Output Homogeneity Back to Base Models
A. Fortier, H. Chen, P. West
EMNLP 2026
Where Do Backdoors Live? A Component-Level Analysis of Backdoor Propagation in Speech Language Models
A. Fortier, T. Thebaud, J. Villalba, N. Dehak, P. Cardinal, P. West
Interspeech 2026
Multi-target backdoor attacks against speaker recognition
A. Fortier, S. Joshi, T. Thebaud, J. Villalba, N. Dehak, P. Cardinal
IEEE ASRU 2025