8d ago · arXiv (AI/CL/LG, keyword-filtered) →
Large language models (LLMs) are increasingly used for text annotation in social science, but standard performance metrics do not explain…