Hi All,
Just wanted to highlight that in the Guidelines video, when we are using prompt engineering to make the response more accurate (the student’s response is supposed to be incorrect), the API returns the correct response (student’s response is incorrect), however when running the notebook, it still insists that the student’s answer is correct.
I have attached the response I got when running the cell
When I run the lab, I get the “incorrect solution”:
Did you change anything in the lab ? Are you trying to run it locally ?
Best,
Mubsi
Hi @Mubsi! No, I am running it in the provided jupyter notebook and didn’t change anything. I have even tried to run it multiple times, but each time I get the same response
Can you share your notebook with me ?
You can download it by clicking on File on the top left and then download. Then you can attach the file here.
Thanks,
Mubsi
Tried to download, it’s not downloading 
I’ll try taking more screenshots and attach them to a post
I guess 3rd time it got it right; I retried now and it gave me the right answer. I wish I’d know what happened the first times I tried…
Thank you @Mubsi for trying to help me with this
I got the same erroneous result from ChatGPT. Attempted to download, but nothing happens.
Second try running and answer is as expected.
Hi @mdesandoli,
I found this to be the best way to download the notebook.
- Open the notebook.
- After the notebook opens up, click
File → Open... - When your workspace opens, click the check box before your notebook file. After it is selected, press
Shutdown. The icon beside the filename should turn from green to gray. - Click the checkbox again and click
Download.
Best,
Mubsi
You can also give additional instructions, in this example it does not work out the (10 * x) correctly. By e.g. providing
- Then simplify the answer.
it will evaluate into 360x + 100,000 which is easier to compare with. I guess this is one of the reasons for trying yourself, changing input etc. It is quite expected that results are not 100% consistent.