Hello! I’m Rebecca Dorn (they/she), a fourth year PhD candidate in Computer Science at the University of Southern California’s Information Science Institute. I'm co-advised by Professors Kristina Lerman and Fred Morstatter. I research fairness in natural language processing and computational social science for social good. I'm particularly interested in understanding how technical tools amplify systems of inequity, and what can be done to mitigate the damage.

Lately, I've been focusing on how language models make sense of dialects used by historically marginalized groups. My recent work found a tendency for LLMs to erroneously flag texts from gender-queer individuals as harmful, with chain-of-thought prompting yielding limited improvement. If any of this sounds interesting to you, let's connect! Feel free to take a look at my academic CV and reach out to me at rdorn[at]usc[dot]edu.

Recent News

Sept 2024 Excited to share that two papers have been accepted for presentation at the EMNLP 2024 Main Conference! OATH-Frames: Characterizing Online Attitudes Towards Homelessness with LLM Assistants and COMMUNITY-CROSS-INSTRUCT: Unsupervised Instruction Generation for Aligning Large Language Models to Online Communities.
July 2024 The paper Harmful Speech Detection by Language Models Exhibits Gender-Queer Dialect Bias has been accepted to EAAMO 2024!
July 2024 Our paper Non-Binary Gender Expression in Online Interactions is accepted to ASONAM 2024 main conference.
June 2024 First day of computer science with USC ISI's Stimulating STEM Summer Program! An honor and a joy to be part of such an incredible team.
June 2024 Passed my qualifying examination for the PhD. I'm now a PhD candidate!
June 2024 Had a lot of fun with Adam Russell as a guest on AI/nsiders podcast. Listen to my personal opinions on the episode "There is no end goal to AI ethics, there will always be something new to mitigate".