Philosophically-flavored questions

Analysis by Charles Rathkopf Last updated: January 2026

[Questions to be written]

Abstract

We evaluate models for in-context scheming using the suite of evals presented in our in-context scheming paper (released December 2024) with the most capable new models.

Citation for this analysis

Charles Rathkopf, “Philosophical Questions in More Capable Models Are Better At In-Context Scheming,” AI Deception Papers, January 2026, [URL]