I am a PhD candidate in computational linguistics at Heinrich-Heine University (HHU) and I work as a doctoral researcher at the DFG-funded project “Coercion and Copredication as Flexible Frame Composition”.

I am interested in language modeling and representation learning. My PhD thesis focuses on how effectively language model representations encode meaning, and how we can improve them to better handle the flexibility of meaning in language.

Topics I have worked on:

• Interpretability of language model representations
• Learning invariant representations through adversarial forgetting to improve task performance
• Exploring the topology of language model representation spaces and its relation to task performance