I think you missed our paper that actually analyzes LLMs' understanding of concepts (I'm referring to "I did not come across many papers that tried to analyze LLM properties").
The name is "Towards Concept-Aware Large Language Models" (by Chen Shani, Jilles Vreeken and Dafna Shahaf).
I think you missed our paper that actually analyzes LLMs' understanding of concepts (I'm referring to "I did not come across many papers that tried to analyze LLM properties").
The name is "Towards Concept-Aware Large Language Models" (by Chen Shani, Jilles Vreeken and Dafna Shahaf).