Evidence of a log scaling law for political persuasion with large language models

Open in new window