From the notebook
Everything I've written so far. Newest first.
- The crossroads: local LLM or cloud APIAug 2026Cloud models are faster and smarter, local models keep your data yours. Thoughts on the tradeoff from someone stuck at that crossroads, and why privacy isn't the argument.
- Persona-vector probes for medical LLMsJun 2026Reading harmfulness, uncertainty, and other behaviors straight off a medical LLM's activations, and the custom probe builder that lets a clinician define new ones on demand.
- From 0.50 to 0.88 AUROC with activation normalizationJun 2026Why normalizing residual-stream activations before a linear probe is the highest-leverage change for behavioral detection in LLMs.
- Sparse autoencoders, liveJun 2026An SAE turns a model's opaque residual stream into a cloud of readable features. Here's how GlassBox uses one as a live interpretability view, and why we don't trust it just on its own.
- What I learned building ExoSeekerNov 2025Lessons from building an AI exoplanet-detection pipeline on NASA Kepler data that hit >90% accuracy and won Best Use of NASA Data.