The gpt-4-1106 preview function call returns thousands of "\\n"

When I use gpt-4-1106-preview to make an api call, I specify it as a function call. The model returns a lot of “\n”, and I have no problem using it on the playground.

There is no requirement in my prompt for the model to output ‘\ \ n’ infinitely. I tried to add the instruction: do not output ‘\ n’, but it was not effective。

1 Like