Projects
Constructive or Conditional? Racial and Gender Bias in LLM Essay Feedback
Systematically identifying differences in LLM essay feedback across race and gender groups
Stanford University CS281
Auditing Racial Bias in LLM-Generated Crime Suspect Descriptions
Systematically identifying differences in LLM-generated suspect descriptions across racial groups
Stanford University CS221
Beeline Reader AI Literacy Platform
An AI-powered, K-12 literacy webpage
Stanford University CS 51 & 52
"Let's Not Talk Politics": Prompt-Based Machine Unlearning for Political Bias
Evaluating the robustness, performance, and strength of prompt-based unlearning for political bias in LLMs
Stanford University CS224N
Quantifying Racial Disproportionality in Fatal Police Shootings Across Contextual Factors
A Bayesian network approach for analyzing fatal police shootings in the US
Stanford University CS109