I am a final-year computer science Ph.D. candidate at University of Toronto in the Natural Language Processing Group and a research scientist intern at Meta AI. I am interested in Large Language Model (LLM) and its applications in Natural Language Processing (NLP).

My current research is focused on 1) invesitgating the principled mechanisms of how Large Language Models solve complex real-world problems, and 2) leveraging the gained insights to design better algorithms to align state-of-the-art LLMs with human values. I am particularly interested in understanding why language models sometimes generate harmful or untruthful content, and this line of research has the potential to enhance the safety and transparency of LLMs.

During my PhD, I also studied computational modeling of creative language generation such as metaphor and idiom.

For more information, you can find my CV here.

Contact: jadeleiyu [at] cs [dot] toronto [dot] edu