LIFEBench/LIFEBench
收藏官方服务：
资源简介：
LIFEBench是一个包含10,800个实例的全面基准，用于评估大型语言模型遵循长度指令的能力，涵盖4个任务类别，包括问题回答、总结、文本生成等，支持英文和中文两种语言。数据集实例的长度约束从16到8192个单词。
LIFEBench is a comprehensive benchmark consisting of 10,800 instances designed to evaluate the ability of large language models (LLMs) to follow length instructions, covering 4 task categories in both English and Chinese, with length constraints ranging from 16 to 8192 words.
提供机构：
LIFEBench
