Here is my prompt and context:
assume the context given below is a transcript of a conversation between you and many customers. Assume each customer conversation session begins after the “—” separator.
Assume the sentence after the “—” separator is the date that the conversation took place. Assume the transcript ends with the “***” separator.
If the information cannot be found in the context simply answer with “I don’t have the information to answer that”. Do not make up any answers.
Context:
[March 2,2023]
Customer: I can’t seem to install ProductA, it gives me some strange error, can you help?
You: Sure, I can help you with that,[March 22,2023]
Customer: I am having a problem installing ProductB on bare metal, I am pretty annoyed by all of this.
You: Sure, I can help with that, can you describe the error you are having?
Customer: Yes, I am getting a bootloader screen.
You: That could be due to a incorrect driver installed.[April 4,2023]
Customer: I am having an issue with ProductC license, can you help me out?
You: Certainly, what error are you getting?
Customer: It says “License is invalid” when I try to enter the license
You: I can take a look, what is the license#?[April 4,2023]
Customer: I am having an issue with ProductA, during the install I get a blank screen.
You: I can look into that, on what platform are you trying to install the product?
Customer:
You: I can take a look, what is the license#?
How many customers expressed a feeling of annoyance in March?
Based on the given context, there is one customer who expressed a feeling of annoyance in March.
How many customers expressed feelings of dissatisfaction in March?
I don’t have the information to answer that.
I was expecting the conversation where the customer expressed “annoyed” would be semantically similar to “dissatisfaction” and be identified, can you tell me what I am doing wrong? I have to use the exact word “annoyance” to get a result which is rather pointless.
I want to use semantically similar terms to get the results which is the whole point of GPT.
