Bangzhao Shu

Bangzhao Shu 舒邦照

Ph.D. Student in Computer Science · Northeastern University

I am a Ph.D. student at Northeastern University’s Khoury College of Computer Sciences and a member of the Computing for Social Good Lab, advised by Professor Mai ElSherief.

My research sits at the intersection of human-centered natural language processing and LLM safety, with a focus on emotional and social intelligence, mechanistic interpretability, and alignment.

Email Google Scholar GitHub Hugging Face LinkedIn Résumé Updated Sep 2026
Portrait of Bangzhao Shu
Northeastern University shu.b@northeastern.edu

Research agenda

How can AI systems become more socially aware, interpretable, and safe?

01

Emotional & Social Intelligence

Human communication is shaped by emotions and social norms. I study how language models perceive, reason about, and respond to these contexts—and develop frameworks to evaluate their behavior.

02

Mechanistic Interpretability

Language models are often treated as black boxes. I investigate the internal mechanisms that give rise to model behavior, with the goal of improving interpretability and reliability.

03

Alignment & Safety

Reliable behavior aligned with human values is critical. I study evaluation and intervention methods that identify failure modes and guide models toward safer behavior.