Marked Personas: Using Natural Language Prompts to Measure Stereotypes in Language Models

Open in new window