Lucas H. McCabe

Data Science Fellow @ LMI. PhD Candidate @ GW GraphLab.

prof_pic.png

I am an interdisciplinary data scientist interested in networks, both neural and social. Broadly, my research investigates topics in natural language processing and computational social science, with a recent emphasis on trustworthiness in AI. Please see my research page for more.

My PhD studies are supervised by H. Howie Huang. I completed my master’s at Johns Hopkins under Tom Woolf. Before that, I studied math and computer science at Rutgers and worked in the Moghe Lab. I am grateful to have been recognized as a DARPA Riser, Luminary Awardee (LMI), and Bernstein Scholar (Institute for Quantitative Biomedicine, formerly “BioMaPS”).

selected research

  1. Preprint
    esa_schema.png
    Estimating Semantic Alphabet Size for LLM Uncertainty Quantification
    L. H. McCabe, R. Melamed, T. Hartvigsen, and H. H. Huang
    arXiv preprint arXiv:2509.14478, Sep 2025
  2. EMNLP (acc.)
    ood-prompts-zipf.png
    Demystifying optimized prompts in language models
    R. Melamed, L. H. McCabe, and H. H. Huang
    arXiv preprint arXiv:2505.02273, May 2025
  3. Appl. Netw. Sci.
    overdose-network-map.png
    Network analysis of U.S. non-fatal opioid-involved overdose journeys, 2018–2023
    L. H. McCabe, N. Masuda, S. Casillas, N. Danneman, A. Alic, and R. Law
    Applied Network Science, Nov 2024
  4. EMNLP
    propane.png
    Prompts have evil twins
    R. Melamed, L. H. McCabe, T. Wakhare, Y. Kim, H. H. Huang, and E. Boix-Adsera
    In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing , Nov 2024