I am trying to do some analysis on the NPS Scores received for a product and its only considering part of the input and also not filtering the ones which do not have additional feedback and I cant figure out a way to solve it
prompt = f"“”
Format from the text delimited by triple backticks \
as below in a table format with html format and remove that have additional feedback as none
Serial Number:
Company Name:
NPS Score:
Additional feedback: {text}
“”"
response = get_completion(prompt)
print(response)
1). Since it’s not filtering out the records with no additional feedback, experiment with rewording that part. You can try making it a separate sentence, experimenting with the wording to make it extra clear you want it to ignore each record where additional feedback is “none”, etc.
2). I’m not sure why it’s only considering part of the input. Maybe you have so many records in the {text} that it is reaching the maximum limit for the number of tokens in the input. One option to consider in this case – since what you’re trying to do could be done relatively easily with just regular code, and since gpt3 is pretty good at writing basic code, you could try asking it to write you some python code to do this. Then you could use the code with any number of text input records. Experiment and see how it works.