So i was using the Curie model to write Product Descriptions, but the output that is generated is short and doesn’t fill up the token quota that I was given in the settings.
I am using the following settings-
Response Length - 120 ( Output generated comes in around 80-90 tokens , sometimes even less)
Temp - 0.80
Top P - 1
Frequency Penalty - 0.15
Presence Penalty - 0.03
Please help, I’ve been stuck at this for a long time .