Lucrezia Laraspata is a Ph.D. Student at the Computational Intelligence Laboratory (CILab), University of Bari Aldo Moro. Her research focuses on Large Language Models and Trustworthy AI, with particular emphasis on mechanistic interpretability and the analysis of internal model representations.
Her work investigates how information related to truthfulness and hallucinations is encoded in the latent representations of Large Language Models, exploring probing and representation engineering techniques to better understand and control model behavior.
Her research also explores the application of language and multimodal models to Human Capital Management, including the analysis and representation of skills, professional profiles and organizational information.