I am broadly interested in how large language models represent and update knowledge, and how those representations shape model behavior, particularly in multi-agent systems, financial NLP, and inference-time adaptation. I welcome all potential collaborators.