In-Context Impersonation Reveals Large Language Models' Strengths and Biases

Re-assign community