GPT4 generates only part of output and just stops - burning the cap limit

Recently, I have huge problems with GPTs (created by me or available from the Store).

Problem 1: The GPT generates a fraction of output and stops mid generating. Which leads to the need of “regenerating” output. Obviously, it counts in the usage limit. So I burn the cap usage limit for literally 4-5 requests… this is awful lately.

Problem 2: GPTs generated by me at one point just respond something totally unrelated from my prompt, it pulls the info from the knowledge uploaded, but the output has nothing to do with the prompt asked.

Anyone has similar issues? Potential solutions?

1 Like