This repository was archived by the owner on Nov 3, 2023. It is now read-only.
Replies: 3 comments 2 replies
-
|
This is another example: |
Beta Was this translation helpful? Give feedback.
0 replies
-
Beta Was this translation helpful? Give feedback.
0 replies
-
|
The model attends to documents using the Fusion-in-Decoder method (FiD); see the original paper for more details of how the model incorporates external documents. Admittedly, sometimes the model does not attend to the documents, for a variety of reasons. Your second example, however, shows some attention (4th highest mountain is part of the retrieved knowledge). |
Beta Was this translation helpful? Give feedback.
2 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment

Uh oh!
There was an error while loading. Please reload this page.
-
Hello team,

i have been experimenting with blenderbot2.0 and internet search and i am trying to understand how the text generation works.
for example i asked the question: "i am doing fine. what is the latest with the terrorist attack at the airport in Kabul?"
it went through online searches where i can see it gather up to date information like:
but the response was: "I'm not sure, I haven't heard anything about it. Do you have a source for that?"
can anyone from the team tell how it works or point me to some documentation i can use to understand this process?
thank you
Juan
Beta Was this translation helpful? Give feedback.
All reactions