Related Posts
Candy is 50% at jewel & target FYI 🥳
Channel 8 news 🐐
Additional Posts in Data & Analytics Consultants
Got messaged by a C3 . ai recruiter. Read that wlb is bad and that the interview process is absurdly long, but the Glassdoor reviews are 4.2 and can't find actual hours worked posted by anyone. How's the culture really? I'd be aiming for DS consulting, something more functional but with DS/ML concepts as my differentiator.
C3.ai, Inc.
Has anyone else begun to resent data science?
today I choose violence

New to Fishbowl?
unlock all discussions on Fishbowl.



What are you asking about?
If Azure is doing api calls to openai, there are several reasons why they could be different.
Azure may have built a system context into their application to tailor or otherwise slightly alter the response you’re getting.
ChatGPT also has a temperature (basically variance) for certain types of questions so even if you ask ChatGPT the same question you’ll get diff types of responses (depending what the question is).
I haven’t dug into the Azure one specifically (it’s on my to-do) but the ChatGPT large language models are basically chatbots (next level chatbots at that) trained on large datasets. The OpenAI version claims to be trained on the entire Internet. I’m not sure what the Azure version is trained on but if it’s a different dataset than the free one, you’ll get different answers.
The challenge for client implementations is going to be the training dataset. Especially in certain industries such as health care, control and vetting of the training dataset is critical because you can’t take a chance on it going off the reservation and telling someone who asked about common cold symptoms that they have gangrene in their foot.
Can you give an example?