A Survey on the Honesty of Large Language Models

Siheng Li, Cheng Yang, Taiqiang Wu +12

Honesty is a fundamental principle for aligning large language models (LLMs) with human values, requiring these models to recognize what they know and don't know and be able to fai…