ecnu-icalk/ELL-StuLife

ELL-StuLife是一个模拟学生在大学期间学习和生活旅程的基准数据集，用于评估AI代理在真实世界交互中的持续成长和自我进化能力。数据集包括三个核心阶段和十个详细子场景，设计围绕从被动到主动、从环境到记忆、从模仿到学习的三个关键转变。

ELL-StuLife is a benchmark dataset simulating a students college journey of learning and life, designed to evaluate the continuous growth and self-evolution capabilities of AI agents in real-world interactions. The dataset includes three core phases and ten detailed sub-scenarios, designed around three key shifts: from passive to proactive, from context to memory, and from imitation to learning.