I’ve been trying to build on GPT 3.5 turbo API; however, I occasionally get token limit errors. When I reset the API, it sometimes works and sometimes does not. I’ve also tried the same input on chatGPT 3.5, which always worked. I’m wondering if anyone else had a similar issue.
I thought it might be due to the output reaching the token limit, but my temperature is also set to zero, so the output should always be identical.