Analyzing Large Language Models' Response to Inductive Instructions with Dual-critique Prompting
Large language models struggle with inductive instructions, prompting the need for enhancements like DUAL-CRITIQUE to improve truthfulness and helpfulness.