Feature Request: An option to parse escaped characters from LLM output.

[sigmareaver]

I'm testing with a Llama 3 derivative, and disclaimer, I haven't yet modified the LlamaTale prompt to follow Llama 3 formatting, but some of the generated output I got from Llama 3 contained \r\n and I think it would be nice to have a config option for handling escaped characters, with perhaps 3 settings:

1. Off
2. Ignore/Discard
3. Convert escape sequences to char (or whatever the standard is)

[sigmareaver]

For what it's worth, I don't believe \r should show up without \n, but \n can show up without \r, so a simple substitution of:
"\\r" = "" and "\\n" = '\n' should work. Forgive my psuedo-code. I know how I'd write this in C, but I'm not very familiar with Python.