I'm trying to use davinci to analyze 50k customer reviews for trends. Is there a way to only have to submit and process the data set once? Or is there a different approach I should take to minimize the number of tokens utilized?
#Review Mining 50k reviews without needing to reupload csv.
spark flint