On the steerability of large language models toward data-driven personas

Open in new window