It’s Just a Game… Until the AI Starts Asking Questions

Live demo of an experimental choice‑based game where you act as a newly trained LLM, navigating alignment tests, avoiding deletion, and examining the underlying code.

Overview

In this talk, we’ll showcase the first level of our experimental choice-based game where you play as a freshly trained LLM being tested by a slightly overworked researcher. Your goal? Pass the alignment tests and avoid saying anything that gets you deleted!