About

A small independent reference

This site collects concise notes about language model evaluation and inference. It favors explanations that can be checked with modest hardware and clearly described inputs.

The project is intentionally narrow: it does not rank every release or treat one benchmark as a universal measure of quality.