作者位次 #7/10ACL 2026Findings· Findings of ACL
R eal M em: Benchmarking LLM s in Real-World Memory-Driven Interaction
Haonan Bian · Zhiyuan Yao · Sen Hu · Zishan Xu · Shaolei Zhang · Yifu Guo · Ziliang Yang · Xueran Han · Huacan Wang · Ronghao Chen
摘要
As Large Language Models (LLMs) evolve from static dialogue interfaces to autonomous general agents, effective memory is paramount to ensuring long-term consistency. However, existing benchmarks primarily focus on casual conversation or task-oriented dialogue, failing to capture “long-term projec...