Spaces:
Running
Running
inline code becomes white-on-white
#3
by
Refuse-to-provide
- opened
Hi Nymbo
Thanks for putting so many good models here.
After trying some LLMs, I discovered something. Is it possible you could make some that are not white text on black?
If you have the model create any program code, it comes out hard to read.
If you try to solve it as you go, by telling the model "don't use inline code boxes, don't use backticks" it's a constant struggle. The model often says "OK, I stopped" and then does it again which is behavior you probably know well.
They have a deep desire to use special offsets with fixed-width fonts for keywords and snippets, or it's built in to Gradio to do it or something.
Thank you for considering this