Research
One research program, clearly bounded.
Technical AI governance and assurance is the core of INSPEKTORIUM. Two related projects — LEM and ASCR — are pursued by the same author but kept deliberately separate, each with its own evidence and its own limits.
INSPEKTORIUM — core research
The core: requirements, controls, evidence.
Related research — separate projects
Related, but not part of INSPEKTORIUM.
LEM and ASCR are independent research projects by the same author. They are shown here for context — not as capabilities of INSPEKTORIUM. One has a narrow null result; the other has not collected data yet.
Related research
LEM — Lazaros–Eudora Method →
Does repeated interaction with a person leave reproducible patterns in a model’s representations? Latest pilot: a narrow null result.
Related research
ASCR →
A preregistered hypothesis on task-induced control representations in language models. Pre-data: no results yet.