I earned my PhD in Computer Science from the University of Arizona, advised by Prof. Mihai Surdeanu. My research interests include Machine Learning, Large Language Models (LLMs), Natural Language Processing, and AI Safety. I am particularly interested in stress-testing the βalignmentβ of LLMs.
My PhD dissertation is the first to identify the issue of data contamination (data leakage) in LLMs, scenarios where training data overlaps with evaluation data. I developed several methods to detect and estimate contamination in fully black-box LLMs. My PhD research received media coverage and earned several awards, including the Outstanding Graduate Scholarship and the Galileo Circle Scholarship.
Previously, I was a research intern at Google Cloud AI Research, Walmart Global Tech, and Harvard Medical School.
Selected Publications
Intent Laundering: AI Safety Datasets Are Not What They Seem
, Marc Wetter
arXiv 2026 | Paper
Blog Post
Media Coverage
Medium Post
YouTube
X Thread
Towards Compute-Optimal Many-Shot In-Context Learning
, Yanfei Chen, Rujun Han, Manan Gandhi, Tianli Yu, Swaroop Mishra, Mihai Surdeanu, Rishabh Agarwal, Chen-Yu Lee, Tomas Pfister
COLM 2025 | Paper
Poster
Memorization in In-Context Learning
, Mihai Surdeanu, Steven Bethard, Eduardo Blanco, Ellen Riloff
arXiv 2025 | Paper
Data Contamination Quiz: A Tool to Detect and Estimate Contamination in Large Language Models
, Mihai Surdeanu
TACL/ACL 2025 | Paper
Poster
Video
Media Coverage
Grading Massive Open Online Courses Using Large Language Models
, Nikhil Garuda, Christopher Impey, Matthew Wenger
COLING 2025 | Paper
Time Travel in LLMs: Tracing Data Contamination in Large Language Models
, Mihai Surdeanu
ICLR 2024 β Spotlight π (notable top 5%)
| Paper
Poster
Video
Media Coverage
Do not Mask Randomly: Effective Domain-Adaptive Pretraining by Masking In-Domain Keywords
, Mihai Surdeanu, Nazgol Tavabi, Ata Kiapour
ACL 2023 RepL4NLP | Paper
Poster
A Natural Language Processing Pipeline to Study Disparities in Cannabis Use and Documentation Among Children and Young Adults: A Survey of 21 Years of Electronic Health Records
Nazgol Tavabi, Marium Raza, Mallika Singh, , Harsev Singh, Grant Hogue, Ata Kiapour
Nature Digital Medicine | Paper
Building Large-Scale Registries from Unstructured Clinical Notes Using a Low-Resource Natural Language Processing Pipeline
Nazgol Tavabi, James Pruneski, , Mallika Singh, Ryan Sanborn, Benton Heyworth, Amir Kimia, Ata Kiapour
Artificial Intelligence in Medicine | Paper
Blog Posts
The AI Safety Illusion: Why Current Safety Datasets Fool Us on Model Safety
, Marc Wetter
Labelbox | Link
Reflections on NeurIPS 2025: Advancing Evaluation and Continual Learning in AI
, Smit Modi, Stepan Tytarenko, Almas Abdibayev, Marc Wetter
Labelbox | Link
Service to the Field
Area Chair: COLM 2026
Reviewer: ICML 2026, AAAI 2026, NeurIPS {2025, 2024}, COLM 2025, ICLR 2025, ACL {2024, 2023}