作者位次 #2/10ACL 2026Findings· Findings of ACL
MDC -Bench: A Multidisciplinary Causal Benchmark Based on Causal Structures for Evaluating Large Language Models
Peng Wang · Yuxiong Yan · Xiao Ding · Kai Xiong · Bibo Cai · Chao Peng · Yutai Hou · Dandan Tu · Bing Qin · Ting Liu
摘要
Existing causal datasets primarily focus on the commonsense domain, where the questions mainly involve simple, single-hop direct causal relationships. When models possess the corresponding knowledge, even if they cannot understand the causal relationships, they can directly arrive at the correct...