Aspiring AI Interpretability Researcher
Luis Miguel
Montoya
Understand to Align
Exploring the journey from physics to AI interpretability research.
Journey Oct 2025
September–October 2025 Recap
Highlights, milestones, and lessons learned from the September–October leg of my AI safety journey.
Project Sep 2025
Towards Disentangling Latent Content and Behavioral Inhibition in Taboo Language Models
Project log for my MATS 9.0 application work probing how latent content and behavioral inhibition interact inside Taboo-fine-tuned Gemma models.
Literature Sep 2025
A Mathematical Framework for Transformer Circuits
Anthropic’s foundational write-up that formalizes how circuits emerge inside transformers and how to reason about them.
Paper