About the Bugs / Site Issues category

Hi,
These kinds of formatting quirks are common across LLMs. Even something as simple as “don’t use em dash” took OpenAI almost a year to fix. Clear model reminders and some trial-and-error with different models often help.