Three papers accepted to EMNLP Main 2025; one paper accepted to ICWSM 2026; one paper accepted to the Actionable Interpretability Workshop @ ICML; gave a talk on detoxification of LLMs using SAEs at the AImpact Center @ UIUC; one paper on small language models for content moderation accepted to NAACL 2025 (Main) as an Oral talk.
Research Experience
Worked as a Research Intern at Adobe Research, mentored by Dr. Apoorv Saxena and Dr. Koyel Mukherjee.
Education
Ph.D. student in Computer Science at the University of Illinois Urbana–Champaign, co-advised by Prof. Hari Sundaram and Prof. Eshwar Chandrasekharan; Undergraduate in Computer Science, Mathematics, and Data Science from the University of Wisconsin-Madison, advised by Prof. Hanbaek Lyu and Prof. Junjie Hu.
Background
Research interests include sociotechnical systems (especially LLMs) and how they shape social interactions. Focuses on understanding model safety, modeling social interaction, and improving outcomes through system design.
Miscellany
Undergraduates with a strong background in ML/NLP and experience with PyTorch are welcome to reach out to agamg2@illinois.edu for research opportunities.