Identity as Attractor: Geometric Evidence for Persistent Agent Architecture in LLM Activation Space

Build a tool that visualizes hidden state attractors to verify if an agent's persona is stable across different prompts.

Suggested repo: attractor-viz

"Peek inside your model's soul to see if your persona is actually holding steady."

Estimated effort: 60h