评分依据：Important safety finding: LLMs fail to grasp user intent beyond explicit content
Beyond Context: Large Language Models' Failure to Grasp Users' Intent
原文： arxiv.org
评分依据：Important safety finding: LLMs fail to grasp user intent beyond explicit content