meander 2023-07-11

I was just thinking: there's an opportunity to improve the utility of Code LLMs for Clojure (such as WizardCoder) by building an instruction set from Clojurians conversations.

I'm very curious what someone that actually knows what they're doing with meander can do with LLM's and meander. Especially with fine-tuning on gpt-3.5-turbo and (announced to become available later) gpt-4 fine-tuning. It should be possible to create some fine-tuning interactions that explain the concepts.