Hello everyone!
I have a question regarding evaluation.
I have created a dataset with questions and added an evaluator to the dataset. Now, lets say i want to run those questions against my agent (and not just against an LLM).
How can this be achieved? Maybe with a sandbox?
Also, is there a way to run an actual multi-step evaluation (preferably against my agent but also against LLM), from what i see i can only "mock" the steps actually run only the latest step.
