Agent Builder: Adding agent trace to eval sets

Hi Team,

I have a few sample agent runs, and I see there is a way to feed the trace back to eval sets.

I provided the feedback in the job section, but I am not able to see anything in the agent evaluation set panel.

I followed the documentation, but I don’t seem to find the option.

@tejaskumar.darji

The option is just below where you find traces

Open agent → you will see agent in the middle and at the bottom you would have execution trail , history and evals..under history you would have the filter and the option to select runtime feedback which shows the same

cheers

Okay yeah, I can see that now under the Definition tab of Agent Builder.

Looks like the documentation needs to be updated with these details. The doc is not very clear about these steps.

Thanks for the details.