Mikyo

Thanks for the feedback Hassan. We'll digest this and fix with a vengeance.

Makes too much sense. Will get something cooking

Cursor [repo=Arize-ai/phoenix] fix the above broken link in the UI. Verify the link is valid and appropriate

Cursor[Arize-ai/phoenix] fix the above broken link

Cursor fix the broken link

hey Andrew S. this is probably not that hard given the state of AI these days. Can you give us a bit more of the type of annotations you'd want to add. Are they basically relevance per document? You need multiple annotations per document? I'll fire off an agent to work - just need to better understand what UX you'd like

Hey Andrew, your request makes a lot of sense. We tend to think of the reference as the area where ground truth is so it's where the evaluators will eventually pull data from for reference based evaluations. But I think making it possible to "chat" with your spans like this makes sense too. I filed an issue. Let me think through the design. It's sorta like you want the prompt variables to be scoped to the entire example so you can do input / output / metadata in your prompt. https://github.com/Arize-ai/phoenix/issues/10912

Thanks for the feedback here. I think I roughly get some of the usability issues but might badly represent them. Copying Srilakshmi C. to help guide our feedback. Keep the feedback coming.

Hey Tristan, HNY. Given today is a company holiday we might be delayed in response. What part of your tracing are you wanting to render in markdown? In general we have support for markdown in the llm messages but depending on the place we might be able to just do it in the UI