I’ve been adding lists of words that I want it to avoid when writing content in my pickaxe, but it’s not really paying attention.
Is there something specific I need to put into the prompt to ensure it avoids those words?
This feature isn’t super great. It’s an underlying part of the OpenAI API that we exposed.
Some users claim it doesn’t work. Some claim it does work well. We actually removed it briefly, but got so many complaints from users who felt it was ‘vital’ to their Pickaxes’ efficacy, that we added it. I’m embarrassed to say I can’t really comment on detail about it 
I’m using the Claude API, and when using Claude directly, it does a really good job of mimicing voice. Is there a way of exploring why it (Claude) doesn’t work as well inside a Pickaxe? 
Can you explain more what you mean? What sort of difference are you noticing between Claude in Pickaxe and Claude elsewhere?
I just tested it again, and it seems to be OK. Thanks
Glad to hear that Claude is working as expected now