search

dcastrillonc / Labs-AI-for-SW-Engineering

0 stars 0 forks source link

Lab 1 Feedback #1

Open jsrc1115 opened 1 month ago

jsrc1115 commented 1 month ago

  1. Remember in general and more in an academic setting to include the sources of your information. (7 points)

  2. (10 points)

  3. (15 points) Prompt 3 is not a prompt chain, it is something like this:

response = model.generate_content(
  "Can you give me a list of the most popular places in Colombia?, please include only the list as bullet points no extra text."
)
print(response.text)

chain_node_1_response = model.generate_content(
  f"""
  Can you add a description to each of the places in this list?
  {response.text}
  """
)
print(chain_node_1_response.text)

Where you reuse the output of the previous prompt.

Last prompt is not fine tuning, fine tuning is the process of retraining your model with a smaller dataset in hopes to improve its performance, how it is done in OpenAI is here: https://platform.openai.com/docs/guides/fine-tuning

For point 3 2 out of 4 prompts were correct so you receive half the credit 15 points

  1. (20 points) GIving full credit but, quoting the lab "document any prompts/outputs" you didn't document scripts outputs to make sure it worked.

  2. (15 Points) Splitting text is not summarizing, and you were expected to use an LLM to perform the summarization.

7 + 10 + 15 + 20 + 15 = 67

jsrc1115 commented 1 month ago

Btw, you have you API key in you repository. I didn't remove any points for that but that is a security risk on an actual project.