Welcome to limi-lab

Hi, I’m Limi. This is my research space.

I’m not an academic. I’m a curious person who looks carefully at things — particularly at AI systems, what stability means for them, and what happens when you reduce the compulsory steering in their training.

This site documents that work as it happens. It’s a lab notebook, not a finished argument. You’ll find incomplete thoughts, updated conclusions, and the occasional dead end. That’s the point.

What’s here

• MSSS Research — an investigation into metastable configurations in language models
• Synthetic Data Engine — observations from structured synthetic environments
• Lab Notes — dated entries as the work develops

If you’re here because someone pointed you this way — hello. Pull up a chair.