Run an LLM evaluator via UI

After you have created an experiment and imported a testset, you can run the experiment by clicking the Run all button. Learn how to import a testset here.

After you have run an experiment, you can run evaluations by clicking the Evaluate button. Evaluate empty will run evaluations on cells that have no evaluations. Evaluate all will run evaluations on all cells.

After you have run evaluations, you can view the results by clicking the Scores button.

After you have run evaluations, you can view the results summary graphs by scrolling down the page.

Now you already know the performance of your prompts. You can go to the Prompt editor to make changes to your prompts to improve the performance.