Matthew W. Shinkle

Matthew W. Shinkle

Postdoctoral Researcher, UC Irvine

I am a postdoctoral researcher with the Relational Cognition Lab at UC Irvine. My current research uses mechanistic interpretability and behavior to understand how LLMs make moral judgments.

I completed my PhD and MS in 2025 from the University of Nevada, Reno, supervised by Mark Lescroart. My graduate research focused on creating computational mappings between deep neural networks and human brain responses, and using these mappings to interpret and influence brain activity. I've also worked as a Research Fellow with PIBBSS and a ML Graduate Research Intern with Pacific Northwest National Laboratory.

In addition to my postdoc at UCI, I'm currently a Visiting Postdoc with the Causality in Cognition Lab at Stanford — if you're in the area and want to chat let me know!

Projects (WIP: more writeups coming soon)

Activation Plateaus in LLMs

Activation Plateaus in LLMs

Tags: AI
Understanding and Controlling the Brain with Feature Visualization

Understanding and Controlling the Brain with Feature Visualization

Tags: NeuroscienceAI
Simulating the Human Brain with Neural Networks

Simulating the Human Brain with Neural Networks

Tags: NeuroscienceAI