About Me

I am a researcher in Natural Language Processing and Vision-Language Models. I am currently a PhD candidate at the Foundations of computational linguistics Lab (colab) at the University of Potsdam, under the supervision of Professor David Schlangen, where I work on linguistic variation and LLMs interpretability.

Prior Work

• Evaluating the visual scene comprehension abilities of state-of-the-art VLMs in a reference expression generation task, curating a semantically minimal COCO-based dataset and applying attention analysis to interpret model behavior (M.Sc. thesis, Utrecht University, under Prof. Albert Gatt)
• Investigating human-inspired progressive alignment and comparative learning frameworks for compositional concept acquisition in VLMs, presented at the CORE Workshop 2024 at Universitat Pompeu Fabra (UPF)
• Exploring ChatGPT’s information-seeking strategies in the 20-Questions game, which earned the Best Evaluation Paper award at INLG-2023 (University of Trento, under Prof. Raffaella Bernardi)