😀 Hello there, I’m Maggie!
I’m currently a PhD student in the UKRI Centre for Doctoral Training in Speech and Language Technologies and their Applications, at the University of Sheffield, UK. My research focuses on understanding idiom processing in Large Language Models, with a broader interest in the intersection of human language processing and computational models. Perhaps along the way, I can uncover insights into how we process language and how language itself works. I feel incredibly fortunate to be supervised by Dr Nafise Moosavi and Prof Aline Villavicencio.
Prior to my PhD, I completed my undergraduate degree in Linguistics from the Lancaster University. It is there, I was introduced to the fascinating field of computational linguistics. This exposure sparked my passion for exploring the intersection between human processing and modelling complexities of human language and cognition, which still very much shapes my research today.
Research Interests
My research interests are typically concentrated on exploring semantics in LLMs, in particular, I’m drawn to: 1) Pretraining, Interpretability, Memorisation in models when encountering idiomatic expressions. 2) Cognition of humans and language models. 3) What is learned by computational models of language?
News
- May 2026: I have just completed a 6-months internship at Epic Games, working with numerous talented colleagues!
- April 2026: Our two papers have been accepted to ACL 2026!
- March 2026: Helped to organise EACL 2026 - catch me in Morocco!
- Nov 2025: Presenting our paper From Input Perception to Predictive Insight: Modeling Model Blind Spots Before They Become Errors in EMNLP 2025, reach out if you want to have a chat.
- July 2025: Presenting our paper Rolling the DICE on Idiomaticity: How LLMs Fail to Grasp Context in ACL 2025 (Oral), see you in Vienna!
- July 2024: I had a great time attending LxMLS 2024, learned a lot and made some amazing friends along the way!
- June 2024: We are running the first ever multimodal task on idiomaticity. See SemEval 2025 - Task 1 for details.
- May 2024: Our paper “Sign of the Times: Evaluating the use of Large Language Models for Idiomaticity Detection” won the Best Paper Award @ MWE 2024! Thanks to my collaborators!
Let’s Connect!
I’m always up for collaboration and discussing ideas. Feel free to drop me an email if you would like to chat.
