Absorbing and Mirroring Personalities of Users of Large Language Models
π Contact Information
Department of Psychology
777 Glades Road
Boca Raton, FL 33431-0991
S.E. Wimberly Library
Sandbox
[email protected]
π Project Leader
Richard Matic
π Collaborators
Jasmine Makkar, Avery Lawrence
π Supervisors
π Project Description
This Ph.D. dissertation is intended to test a novel way to absorb the personality of users of large language models (LLMβs). Dramatis persona is a construct in creative writing which emobodies a characterβs personality in just a few sentences and can carry that character through dramatic arcs in multiple created or LLM-generated stories. This project will attempt to extract a dramatis persona of a user based on the userβs dialog and personal descriptions. That dramatis persona will then be used to synthesize a matching (or mirrored) personality of the LLM. The method will be compared to other approaches to personality mirroring, including Big Five.