LLM

These notes are about running models on my own hardware. The long comparison is still running, so there are no results to brag about yet.

Pinned

Learning log is open

The learning log is open. The full 15-task run is still going, and accuracy comes after that.

A careful test takes days. That's frustrating.