1 paper
Praveen Kumar Myakala, Manan Agrawal, Rahul Manche
LLMs are increasingly used as long-running conversational agents, yet every major benchmark evaluating their memory treats user information as static facts to be stored and retriev…