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

Persuasiveness of generative AI models