Is Your LLM Really Mastering the Concept? A Multi-Agent Benchmark

Shuhang Xu, Weijian Deng, Yixuan Zhou +1

Concepts serve as fundamental abstractions that support human reasoning and categorization. However, it remains unclear whether large language models truly capture such conceptual…