search

microsoft / autogen

A programming framework for agentic AI 🤖
https://microsoft.github.io/autogen/
Creative Commons Attribution 4.0 International
30.53k stars 4.44k forks source link

[Feature Request]: Implement SAFE Agent for fact-checking #2185

Open rokosbasilisk opened 5 months ago

rokosbasilisk commented 5 months ago

https://github.com/google-deepmind/long-form-factuality/tree/main https://arxiv.org/pdf/2403.18802.pdf SAFE is a method to split longform answers and check factuality, ConverseableAgent can be extented to implement this.

rokosbasilisk commented 5 months ago

I am currently working on this, if this is not in-line with autogen's priorities the issue can be closed.

WaelKarkoub commented 5 months ago

@rokosbasilisk can it be implemented as a capability?

Josephrp commented 5 months ago

my impression is that you shoulld build what you like then contribute it upstream as you see fit, a good start is a notebook demo , for a capability , it could be interesting but a lot more work than a simpler demonstration.

gagb commented 1 week ago

@rokosbasilisk thank you! Let us know if there are any updates.