SUMMER RESEARCH FELLOWSHIP 2024

The XLab 2024 Summer Research Fellowship (SRF) ran from June 10 to August 16. We had a cohort of 20 fellows undertaking independent research projects spanning a range of risk areas, each with mentorship from a domain expert. Seven fellows worked on nuclear security projects, eight on technical AI governance and safety, two on biosecurity, and two on climate change. The experience levels of the cohort ranged from second-year undergraduates to incoming graduate students. The final cohort was selected from more than 250 applicants.

A typical week’s programming consisted of a workshop and discussion, a peer review session, a cohort social event, and 1-1 meetings with the program manager and the student’s project mentor.

Guest speakers included:

John Mecklin, Editor-in-Chief, Bulletin of the Atomic Scientists
Kevin Frazier, Professor and Lawfare Fellow
Nicole Kikendall, Principal Biologist, Argonne National Laboratory
Max Lamparth, Postdoctoral Fellow, Stanford Existential Risk Initiative

2024 FELLOWS AND THEIR PROJECTS

COLE SALVADOR

Computer Science and Statistics, Harvard University ‘26
Mentor: Keller Scholl, RAND

Certified Safe: A Schematic for Approval Regulation of Frontier AI

AUDREY WECKWERTH

BA/MA Global Studies and International Relations, University of Chicago ‘25
Mentor: Christopher Twomey, Naval Postgraduate School

Strait to Catastrophe: The Role of Perceptual Factors in US-China-Taiwan Defense Posture and Nuclear Escalation

AMINA ANOWARA

Chemical and Biological Engineering, Princeton University ‘25
Mentor: Junhong Chen, University of Chicago; Aaron Shi, University of Chicago

Integration of Remote Floating Gate FET-based Biosensor with Millifluidic Systems For Point-Of-Care Diagnostics

ARYAN SHRIVASTAVA

Mathematics, University of Chicago ‘26
Mentor: Jessica Hullman, Northwestern University

Response Inconsistency of Large Language Models in High-Stakes Military Decision Making

RHEA KANUPARTHI

Political Science and Mathematics, University of Chicago ‘27
Mentor: Ankit Panda, Carnegie Endowment for International Peace

Nukes and Ladders: India and Pakistan’s Nuclear Evolution

SIMON SISKEL

Political Science and Mathematics, University of Chicago ‘25
Mentor: Paul Bracken, Yale University

Deterrence and Deescalation: U.S. Strategies for a Nuclear-Capable Iran

ANASTASIA WEI

Physics and Data Science, Northwestern University ‘25
Mentor: Kiho Park, University of Chicago

Locating Semantic Understanding: A Study of Semantically Similar Sentences in Large Language Models

SANYU RAJAKUMAR

Computational Biology, Brown University ‘24
Mentor: Rudolf Laine, Oxford University

Eliciting Situational Awareness in Large Language Models Through Fine-Tuning

SOPHIA LLOYD-GEORGE

Physics and Philosophy, Computer Science, Brown University ‘25
Mentor: Bo Li, UIUC; Stephen Casper, MIT

Team Project: Breaking the Circuit Breaker: When does Circuit Breaking Safety Unlearning Fail?

KEVIN LI

MA, Committee on International Relations, University of Chicago ‘24
Mentor: Charles Glaser, MIT

Tripolarity, Resolve, and Nuclear Risks

LAVINIA CAVALET

BA/MA Economics and International Relations, University of Chicago ‘25
Mentor: Austin Carson, University of Chicago

To Arm or Not to Arm: an Arctic Dilemma

JAMES ROSENBERGER

Computational and Applied Mathematics, Statistics, University of Chicago ‘26
Mentor: Mihai Anitescu, University of Chicago

Measuring Community Significance and State Vulnerability After Nuclear Winter

HELENA TRAN

Mathematics, University of California Irvine ‘25
Mentor: Olli Järviniemi, University of Helsinki

Situational Awareness and Consistency in Language Model Deception: a Simulated AI Company Assistant

ROBERT SHIN

Physics and Statistics, University of Chicago ‘27
Mentor: Carter Teplica, New York University

Investigating Fourier Representations and Learned Algorithms within Grokked Transformers

MELODY GUI

Computer Science and Philosophy, University of Southern California ‘25
Mentor: Bo Li, UIUC; Stephen Casper, MIT

Team Project: Breaking the Circuit Breaker: When does Circuit Breaking Safety Unlearning Fail?

ABHAY KATOCH

Biotechnology, Queen’s University at Kingston ‘25
Mentor: Jonathan Ozik, University of Chicago/Argonne

Modeling the Optimal Distribution of Wastewater-Based Surveillance Techniques

JIBANG WU

PhD, Computer Science, University of Chicago
Mentor: Haifeng Xu, University of Chicago

Grounded Persuasive Language Generation for Automated Marketing

AUSTIN COFFEY

MPhil in Policy Analysis, RAND Pardee ‘26
Mentor: James Steinberg, former U.S. Deputy Secretary of State, Johns Hopkins University

Managing Conflict in the South China Sea

JULIEN BENCHEK

Political Science, Economics, and Physics, University of Chicago ‘27

Large Language Models Exhibit Cognitive Biases in Interactive Stock Trading Scenarios