Topped M output issues | OpenRouter | Page 1

dusky knot Nov 21, 2023, 4:10 PM

#

Playing around with Toppy M and keep getting finish_reason: length.

Prompt tokens are only about 100 but it won't give me more than 65 or 70 tokens in the answer.

So where I'm expecting a lot more of a response I'm getting only a couple sentences.

rugged mist Nov 21, 2023, 9:06 PM

#

what is your max_tokens and what kind of prompt are you sending?

rugged mist Nov 21, 2023, 9:51 PM

#

working on extending the max output here too

dusky knot Nov 21, 2023, 9:53 PM

#

I don't have max tokens set. I was unaware that was an available parameter.

dusky knot Nov 21, 2023, 9:55 PM

#

So can I add the parameter to the json the same way I would an openai call?

#Topped M output issues