Tom Goldstein's Lab at University of Maryland, College Park

AI & ML interests

AI security & privacy, algorithmic bias, foundations of ML

Recent Activity

JonasGeiping  new activity 9 days ago

tomg-group-umd/huginn-dataset:Improve Huginn Dataset card: Add paper/code links, sample usage, and update formatting

montehoover  updated a dataset about 1 month ago

tomg-group-umd/DynaBench

JonasGeiping  authored a paper about 1 month ago

Strategic Dishonesty Can Undermine AI Safety Evaluations of Frontier LLM

View all activity

Papers

Gemstones: A Model Suite for Multi-Faceted Scaling Laws

Transformers Can Do Arithmetic with the Right Embeddings

View all Papers

Team members 29