Get the latest tech news

Something weird is happening with LLMs and Chess

are they good or bad?

I only ran 10 trials since AI companies have inexplicable neglected to send me free API keys and this was costing The Automator money. Theory 1: Base models at sufficient scale can play chess, but instruction tuning destroys it. If you want to know exactly how I did things, here are some words: I ran all the open models (anything not from OpenAI, meaning anything that doesn’t start with gpt or o1) myself using Q5_K_M quantization, whatever that is.

Get the Android app