Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

feat: render markdown in input/output fields #2109

Closed
marcklingen opened this issue May 20, 2024 Discussed in #2097 · 4 comments
Closed

feat: render markdown in input/output fields #2109

marcklingen opened this issue May 20, 2024 Discussed in #2097 · 4 comments
Assignees

Comments

@marcklingen
Copy link
Member

Discussed in https://github.com/orgs/langfuse/discussions/2097

Originally posted by hburrichter May 18, 2024

Describe the feature or potential improvement

It would be great to see the rendered Markdown of a trace input/output or other observations in the Langfuse UI.

This would greatly improve the readability of Markdown elements like tables and images. Currently, we need to copy and open image links manually to view and check them.

Perhaps a "Markdown" option could be added to the existing "Pretty | Json" button group in Langfuse?

Additional information

No response

@marcklingen
Copy link
Member Author

great idea, @hburrichter do you use ChatML (rendered in "pretty" mode) or plain text?

@hburrichter
Copy link

Thanks Marc for the fast response!

The input/output of our traces are just two strings, the user and assistant message content. So we not even have this option/button group (Pretty | Json) in the trace view.

Within our traces, we use Pretty/ChatML for quickly checking the list of messages, and JSON for more detailed evaluations. For instance, we pass messages and a tools as one large JSON as the input of certain observations and the ChatML view shows us the nicely formatted list of messages and the JSON view lets us check the tool definitions that were passed alongside the messages to GPT. This works really well for us.

@marcklingen
Copy link
Member Author

Thanks Marc for the fast response!

The input/output of our traces are just two strings, the user and assistant message content. So we not even have this option/button group (Pretty | Json) in the trace view.

Within our traces, we use Pretty/ChatML for quickly checking the list of messages, and JSON for more detailed evaluations. For instance, we pass messages and a tools as one large JSON as the input of certain observations and the ChatML view shows us the nicely formatted list of messages and the JSON view lets us check the tool definitions that were passed alongside the messages to GPT. This works really well for us.

Thanks, super helpful!

@marliessophie
Copy link
Member

Markdown is now supported in Pretty mode for trace input/output. See changelog post

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants