pull down to refresh

I have to refactor a project I've been working on and pay for more LLM API calls because I realized I was calculating perplexity wrong.