作者位次 #4/9ACL 2026Findings· Findings of ACL
J udge A gent: Beyond Static Benchmarks for Knowledge-Driven and Dynamic LLM Evaluation
Zhichao Shi · Xuhui Jiang · Chengjin Xu · Cangli Yao · Shengjie Ma · Yinghan Shen · Zixuan Li · Jian Guo · Yuanzhuo Wang
摘要
Current evaluation methods for large language models (LLMs) primarily rely on static benchmarks, presenting two major challenges: limited knowledge coverage and fixed difficulties that mismatch with the evaluated LLMs. These limitations lead to superficial assessments of LLM knowledge, thereby im...