Good results fine tuning a local LLM like Qwen 3:0.6B to categorize questionsteachmecoolstuff.com54 pointsdev-experiments6 hours ago