Finetuning an LLM to write docs like its 1995
News Source : Passo.uno
News Summary
- To train a local model to write like a technical writer from the 90s, one needs tons of written sources.
- Bitsavers is a website that collects and scans old computer manuals and brochures.
- I split the sanitized text into training examples on paragraph and section boundaries.
- I then used a cheap and fast model through OpenRouter to classify each paragraph as either “keep” or “drop” based on its intelligibility.
- Compared to full training, fine-tuning doesn’t require a massive amount of data, so it’s cheaper.
In my predictions for 2030 I wrote that tech writers would be using specialized LLMs, running locally on powerful hardware.
Never miss a story from us, subscribe to our newsletter