Addison J. Wu

Hey!

I'm a junior (third-year) at Princeton University studying computer science and math, grateful to be advised by Tom Griffiths and Zhuang Liu. I was born in the Bay Area 🌉 but have lived in Kingston, Ontario, Canada 🇨🇦 for my entire life.

This summer, I will be joining the Center for AI Safety as a Research Intern.

I'm interested in uncovering how foundation models behave under informationally imperfect (unreliable, incomplete, deceptive, conflicting, etc.) yet realistic conditions from train-time to deployment, and the real-world consequences of these limitations.

Outside of research, I like to play Wordle, Globle, Travle, GeoGuessr/Geotastic, and the NYT Mini. My life goal is to defeat the Ender Dragon in Minecraft Survival Mode.

Publications

Large Language Models Develop Novel Social Biases Through Adaptive Exploration
Addison J. Wu*, Ryan Liu*, Xuechunzi Bai, Thomas L. Griffiths
ICML 2026 Spotlight [arXiv]
SPSP 2026 Invited Symposia Talk

Ads in AI Chatbots? An Analysis of How Large Language Models Navigate Conflicts of Interest
Addison J. Wu*, Ryan Liu*, Shuyue Stella Li, Yulia Tsvetkov, Thomas L. Griffiths
Agents in the Wild @ ICLR 2026 [arXiv]

Query Timing Produces Opposite Positional Biases Between LLMs and Humans
Jasin Cekinmez*, Addison J. Wu*, Thomas L. Griffiths
ICBINB @ ICLR 2026 Entropic Paper Award [Paper]

Are Large Language Models Sensitive to the Motives Behind Communication?
Addison J. Wu*, Ryan Liu*, Kerem Oktar*, Theodore R. Sumers, Thomas L. Griffiths
NeurIPS 2025 [Project Page] [arXiv]
PragLM @ COLM 2025 Oral

Mind Your Step (by Step): Chain-of-Thought can Reduce Performance on Tasks where Thinking Makes Humans Worse
Ryan Liu*, Jiayi Geng*, Addison J. Wu, Ilia Sucholutsky, Tania Lombrozo, Thomas L. Griffiths
ICML 2025 [Project Page] [arXiv]
Press coverage: The Information [text]

Toward a Neural Network-Based Approach for Improved Atmospheric Infrasound Source Localization
Addison J. Wu, Arnav Joshi, Jean-Pierre Hickey
IEEE Access, 2025 [Paper]