Daniel KingSenior Project/Solution Manager| NorlysKøbenhavn, Denmark
How do I secure/safeguard that my Generative AI application is not losing its own “train of thought” from the start of the dialog to the end of the dialog, especially where the responses start becoming more complex? Saving Changes...
It comes down to the feedback you provide through the dialog. If you start to detect it is following the wrong breadcrumbs, you need to guide it back by providing constraints or additional information as well as evaluating the quality of the responses it gives you.
Turn the answer into an outline form. I noticed that if the AI tool is only referencing 1 or 2 articles, the training dataset is so small that it starts to go off on tangents about other things mentioned in the article rather than the questions asked. Saving Changes...
Sergio Luis ConteHelping to create solutions for everyone| Worldwide based OrganizationsBuenos Aires, Argentina
I can write a lot about this topic. No enough room here. My recommendation search into the internet about prompt engineering. You will find all you need there. Saving Changes...
Daniel KingSenior Project/Solution Manager| NorlysKøbenhavn, Denmark
Thank you Sergio for responding back, and I really appreciate your patience with my questions which are fairly broad.
I am fortunate to have you and the many other projectmanagement.com participants, that when possible can share what they have seen and experienced.
At my organization, we do have the Generative AI suppliers participating/demonstrating and sharing recommendations in our "lunch&learn" sessions - e.g., today we had Google on Innovation, which was good..
I have also found some webinars fairly inspirational, as some articles that have been shared here and from other sources - and there are countless articles today - OMG.
I’m however missing the "project manager actor" in use case scenarios in most discussion being addressed at this time. Thanks again, I'm looking ;-)
...
1 reply by Sergio Luis Conte
Feb 16, 2024 5:09 AM
Sergio Luis Conte
...
You are welcome. This and linkedin are the only spaces where I am active because I learn a lot from people posts and comments. So, my gratitude is with you for allow me to interact with you and others. Today I am working on one of these suppliers (perhaps the biggest one in this matter and the most advance on it just because one of the "fathers" of this revolution is working there from long time ago). Just take into account that to implement generative AI at organizational level, demands, in most of the case, create a new structure. Why? Because the organization needs to validate the outputs in all sense: if they are "true" or not, data secutiry and data protection, and lot of other things. Because of that, to use AI in general and genAI in particular is like any other type of projects except in this case I fully recommend to find people that has real expertisse on the matter.
Saving Changes...
Sergio Luis ConteHelping to create solutions for everyone| Worldwide based OrganizationsBuenos Aires, Argentina
Feb 15, 2024 8:01 AM
Replying to Daniel King
...
Thank you Sergio for responding back, and I really appreciate your patience with my questions which are fairly broad.
I am fortunate to have you and the many other projectmanagement.com participants, that when possible can share what they have seen and experienced.
At my organization, we do have the Generative AI suppliers participating/demonstrating and sharing recommendations in our "lunch&learn" sessions - e.g., today we had Google on Innovation, which was good..
I have also found some webinars fairly inspirational, as some articles that have been shared here and from other sources - and there are countless articles today - OMG.
I’m however missing the "project manager actor" in use case scenarios in most discussion being addressed at this time. Thanks again, I'm looking ;-)
You are welcome. This and linkedin are the only spaces where I am active because I learn a lot from people posts and comments. So, my gratitude is with you for allow me to interact with you and others. Today I am working on one of these suppliers (perhaps the biggest one in this matter and the most advance on it just because one of the "fathers" of this revolution is working there from long time ago). Just take into account that to implement generative AI at organizational level, demands, in most of the case, create a new structure. Why? Because the organization needs to validate the outputs in all sense: if they are "true" or not, data secutiry and data protection, and lot of other things. Because of that, to use AI in general and genAI in particular is like any other type of projects except in this case I fully recommend to find people that has real expertisse on the matter.
...
1 reply by Daniel King
Feb 19, 2024 9:08 AM
Daniel King
...
Thanks again and I fully agree Sergio with your assessment 👍
Saving Changes...
Daniel KingSenior Project/Solution Manager| NorlysKøbenhavn, Denmark
Feb 16, 2024 5:09 AM
Replying to Sergio Luis Conte
...
You are welcome. This and linkedin are the only spaces where I am active because I learn a lot from people posts and comments. So, my gratitude is with you for allow me to interact with you and others. Today I am working on one of these suppliers (perhaps the biggest one in this matter and the most advance on it just because one of the "fathers" of this revolution is working there from long time ago). Just take into account that to implement generative AI at organizational level, demands, in most of the case, create a new structure. Why? Because the organization needs to validate the outputs in all sense: if they are "true" or not, data secutiry and data protection, and lot of other things. Because of that, to use AI in general and genAI in particular is like any other type of projects except in this case I fully recommend to find people that has real expertisse on the matter.
Thanks again and I fully agree Sergio with your assessment 👍 Saving Changes...
