Running Evaluators in the Playground

[mmabrouk]

Run evaluators directly in the playground to get immediate quality feedback on prompt changes. Instead of switching to a separate evaluation workflow, you can evaluate outputs inline as you iterate on prompts. Scores, pass/fail results, and evaluator reasoning appear right next to the LLM response.

This tightens the feedback loop between prompt engineering and evaluation. You can catch regressions and validate improvements in real time.