Assistant Professor, University of California, San Diego
2 papers at NeurIPS 2025
A principled and scalable method for model-agnostic data valuation
Inference-time approach to mitigate verbatim generation in large language models
