My background is in mathematics, philosophy, and machine learning. I am interested in applying ideas from mechanistic interpretability and causal inference in neuroscience and the natural sciences.

Background

B.S. Mathematics, Philosophy — University of Massachusetts Amherst (2016–2020)

M.S. Computer Science — University of Massachusetts Amherst (2020–2022)

Machine Learning Engineer — Swarm Labs (2023–2024)

Machine Learning Engineer — Eluve Inc (2024–2026)

Visiting Researcher — University of Edinburgh, School of Informatics (2026–present)

Selected Publications

• Rajarshi Das, Ameya Godbole, Ankita Naik, Elliot Tower, Manzil Zaheer, Hannaneh Hajishirzi, Robin Jia, Andrew McCallum. Knowledge Base Question Answering by Case-based Reasoning over Subgraphs. ICML 2022. [Proceedings]

Preprints

• Elliot Tower. Mechanistic Validity: A Theory of Validity for Mechanistic Claims. 2026. [Zenodo]

• Elliot Tower. Bracket Norm Identifies Causally Important Brain Regions From Population Geometry. 2026. [Zenodo]

• Elliot Tower. Direction Instability Predicts Cross-Cell-Line Drug Mechanism Transport in LINCS L1000. 2026. [Zenodo]

Current Research

1. Factorized circuits

Decomposing pretrained transformer weights into shared factor banks with sparse selectors to analyze information flow through the residual stream.

2. Geometric methods for causal inference in the sciences

Grassmannian and curvature-based methods for testing when causal and mechanistic claims transport across populations, instruments, and domains.

3. RNA therapeutic target selection

Computational tools for identifying allele-selective antisense oligonucleotide target sites in repeat-expansion diseases, combining physics-based structure prediction with knowledge-graph reasoning.