Sentiment analysis with few-shot prompting
You're working on market research and your goal is to use few-shot prompting to perform sentiment analysis on customer reviews. You are assigning a number for a given customers conversation: -1 if the sentiment is negative, 1 if positive. You provide the following examples as previous conversations for the model to learn from.
- The product quality exceeded my expectations -> 1
- I had a terrible experience with this product's customer service -> -1
The OpenAI package has been pre-loaded for you.
This exercise is part of the course
Prompt Engineering with the OpenAI API
Exercise instructions
- Provide the examples as previous conversations assigning the text as context for the
userrole and the number as context for theassistantrole. - Provide the following text for the model to classify and use the appropriate role:
The price of the product is really fair given its features.
Hands-on interactive exercise
Have a go at this exercise by completing this sample code.
client = OpenAI(api_key="")
response = client.chat.completions.create(
model = "gpt-4o-mini",
# Provide the examples as previous conversations
messages = [{"role": "____", "content": "____"},
{"role": "____", "content": "____"},
{"role": "____", "content": "____"},
{"role": "____", "content": "____"},
# Provide the text for the model to classify
{"role": "____", "content": "____"}
],
temperature = 0
)
print(response.choices[0].message.content)