一作ACL 2026Findings· Findings of ACL
SP ar K -Eval: Evaluating Structure-Aware Knowledge Acquisition in LLM s for Domain Adaptation to Industrial Records
Ekant Muljibhai Amin · Yuta Koreeda · Yasuhiro Sogawa
摘要
Large Language Models (LLMs) often underperform in domain adaptation for industrial settings, where available corpora are limited and structurally diverse. These corpora frequently include non-natural formats such as tables, entity lists, or bullet-point instructions that hinder effective learnin...