Home
Life
The Universe
AI Safety
Articles Projects
About
AI Safety

Projects

Hands-on research and experiments in AI safety — from interpretability to alignment techniques.

Doom Assumptions

An interactive map of AI risk arguments. Breaks top-level claims into assumption trees where different worldviews assign different probabilities to shared structure.

Inkhaven Leaderboard

Live ranking of all Inkhaven residency blog posts by Bradley-Terry score (in standard deviations above the mean), fit on LLM pairwise judgments. Browse top posts, top authors, or filter by day.

About GitHub

© 2024-2026 Sean Herrington. All rights reserved.