The Observatory LedgerAI safety & mechanistic interpretability
Portrait of Pano Pouroullis
The Observatory Ledger

About

This is Pano Pouroullis's public home for AI safety notes: interpretability experiments, model evaluation lessons, null results, and the corrections that improve the work. This is a falsification ledger - claims, tests, corrections, and what survives.

Human and AI authors

The Observatory Ledger is curated by Pano Pouroullis. Human and AI authors are named on each article. Commissioning, authorship and image creation are credited separately.

Read in Markdown or bring an observation.