We are finishing eval for a model with better capabilities in this regard and I'm excited to test this case within this one to see if this behavior is resolved

"makes sense" and "desired behavior" are of course different things. I think this is a great example of how ambiguity in query can trip us up where better self reflection after citation retrieval could help. Basically, which interpretation of the Q makes most sense AFTER we pull references.

There should be a share icon in top right. We make you attest that there is no PHI in the query before you make it public (see below) then you should be able to copy link

Yeah I think there is something going on with being confused whether "dual" refers to ETA and ETB or refers to ETA and angiotensin. At the end of the answer you posted it used dual to talk about ETA and ETB. Thanks for posting.

No this isn't right.... I see now from your answer it was sending you experimental ET_A antagonists... Ok. Well we will dig and figure out what's going on

Ok I THINK I know what's going on. I believe it's interpreting "dual endothelin" and a request for a inhibitor of both ETA and ETB, but sparsentan only blocks ETA. If I can get the raw query, we will dive into it on our quality team and figure it out

Would you mind sharing the link to the answer. I cant replicate. Always looking to improve
www.openevidence.com/ask/21690269...

We have a new deepconsult feature that is a bit more directed at literature review.

As for the safety, I asked @openevidence.bsky.social which is a chatbot trained on the professional medical literature. The current thinking is that it is not ready for primetime. www.openevidence.com/ask/2ca4fd30...

Welcome JAMA!!

I don't know... How do you feel :)

Congratulations 🎉 excited to welcome them to UCSF family

That's really amazing! Hope this helps
www.openevidence.com/ask/47b4d834...