prompt 1 [논문 리뷰] Thinking Fair and Slow: On the Efficacy of Structured Prompts for Debiasing Language Models Sep 4, 2025