nprofile1qy2hwumn8ghj7un9d3shjtnddaehgu3wwp6kyqpq8c7wjmr8txk9u3xzrxl5rsx8mpt4dr84nyluufn4qg4x9xnar52qr7gqp6 (nprofile…gqp6)
Just as a point of reference, data science folks call their result validation step "evals" and others are (slowly IMHO) building evals for specific domains.