Ivory Yang

ivory.yang.gr [at] dartmouth.edu

Hello! I am a Master’s student in Computer Science at Dartmouth, where I am part of the Minds, Machines and Society lab led by Prof. Soroush Vosoughi. My research focuses on Natural Language Processing and AI Alignment, particularly in the detection of persuasion and enhancement of LLM defenses against persuasive attempts.

In what feels like a different lifetime, I received my Bachelor’s of Business Administration and Bachelor’s of Science in Cognitive Science from the University of Michigan. I worked briefly in finance and start-ups in San Francisco, where I was inspired by the vibrant tech scene to take Computer Science courses at Harvard and Stanford. Eventually, I made the leap of faith to formally return to school.

Given my unconventional journey, I am always excited to connect with others and listen to different perspectives. Feel free to reach out if you would like to collaborate!

news

May 16, 2024	My very first paper is accepted at ACL’24 main conference! Grateful to my first author, Yuxin, for the collaboration opportunity
Mar 13, 2024	Accepted to the Spring 2024 cohort for Supervised Program for Alignment Research (SPAR), organized by UC Berkeley!
Jan 07, 2024	Excited to join the Minds, Machines and Society lab led by Professor Vosoughi

selected publications

ACL

MentalManip: A Dataset For Fine-grained Analysis of Mental Manipulation in Conversations

Yuxin Wang, Ivory Yang, Saeed Hassanpour, and Soroush Vosoughi

ACL (oral presentation, main conference), 2024

Abs arXiv

Mental manipulation, a significant form of abuse in interpersonal conversations, presents a challenge to identify due to its context-dependent and often subtle nature. The detection of manipulative language is essential for protecting potential victims, yet the field of Natural Language Processing (NLP) currently faces a scarcity of resources and research on this topic. Our study addresses this gap by introducing a new dataset, named MentalManip, which consists of 4,000 annotated movie dialogues. This dataset enables a comprehensive analysis of mental manipulation, pinpointing both the techniques utilized for manipulation and the vulnerabilities targeted in victims. Our research further explores the effectiveness of leading-edge models in recognizing manipulative dialogue and its components through a series of experiments with various configurations. The results demonstrate that these models inadequately identify and categorize manipulative content. Attempts to improve their performance by fine-tuning with existing datasets on mental health and toxicity have not overcome these limitations. We anticipate that MentalManip will stimulate further research, leading to progress in both understanding and mitigating the impact of mental manipulation in conversations.