When using gpt 4 (not sure about the others) sometimes the response doesn't get to finish, probably because of the length limit. It starts a new sentence, then hits the limit and stops. Example:
"Yes, I think pineapple on pizza is gross. I mean, whose idea was it anyway? It's like"
Maybe the output length in tokens limit shouldn't be there and instead we should limit the resposne only by prompting?
When using gpt 4 (not sure about the others) sometimes the response doesn't get to finish, probably because of the length limit. It starts a new sentence, then hits the limit and stops. Example:
"Yes, I think pineapple on pizza is gross. I mean, whose idea was it anyway? It's like"
Maybe the output length in tokens limit shouldn't be there and instead we should limit the resposne only by prompting?