User:Pepa Atanasova

Dr. Pepa Atanasova is a tenure-track Assistant Professor in the AI Section at the University of Copenhagen, where she co-leads the NLP group CopeNLU. Her research focuses on making language models trustworthy, transparent, and accountable, investigating the internal mechanisms behind model outputs to steer them toward safer, more factual behaviour. She co-directs an interdisciplinary Trustworthy AI programme and leads a synergy project on trustworthy human-AI interaction, with particular attention to emergent deception in AI-mediated communication.