作者位次 #5/13ACL 2026Findings· Findings of ACL
Sentient Agent as a Judge: Evaluating Higher-Order Social Cognition in Large Language Models
Bang Zhang · Ruotian Ma · Qingxuan Jiang · Peisong Wang · Jiaqi Chen · Zheng Xie · Xingyu Chen · Yue Wang · Fanghua Ye · Jian Li · Yifan Yang · Zhaopeng Tu · Xiaolong Li
摘要
Assessing how well a large language model (LLM) understands human, rather than merely text, remains an open challenge.To bridge the gap, we introduce Sentient Agent as a Judge(SAGE), an automated evaluation framework that measures an LLM’s higher-order social cognition.SAGE instantiates a “Sentie...