Research report: Explaining alignment between LLM classifications of political arguments and human judgments Bioengineer.org