Talking to chatbots with curiosity in the way of "What will it do here?" is an ablation for evaluations/benchmarking.