Abstract
"本文聚焦于“叙实性推理”任务,即判断语言中事件真实性的语义理解能力。该任务不依赖外部知识,而基于语言结构本身进行推理,对当前大语言模型(LLMs)提出挑战。为解决模型在叙实性漂移、多义词处理等方面的不足,作者提出一种结合RAG(检索增强生成)与谓词相似性的方法,构建了一个融合参数化与非参数化知识的叙实性检测智能体系统。该系统通过分步提示与知识库支持,实现了更高的一致性、准确性与可解释性,在评测任务中取得了0.9240的稳健表现。"
- Anthology ID:
- 2025.ccl-2.18
- Volume:
- Proceedings of the 24th China National Conference on Computational Linguistics (CCL 2025)
- Month:
- August
- Year:
- 2025
- Address:
- Jinan, China
- Editors:
- Hongfei Lin, Bin Li, Hongye Tan
- Venue:
- CCL
- SIG:
- Publisher:
- Chinese Information Processing Society of China
- Note:
- Pages:
- 152–156
- Language:
- URL:
- https://aclanthology.org/2025.ccl-2.18/
- DOI:
- Bibkey:
- Cite (ACL):
- Yu Wang, Yang Qian, Ke Liang, Yiheng Yang, Zhai Yu, and Chu-Ren Huang. 2025. CCL25-Eval任务四系统报告:基于RAG与谓词相似性方法的叙实性检测智能体. In Proceedings of the 24th China National Conference on Computational Linguistics (CCL 2025), pages 152–156, Jinan, China. Chinese Information Processing Society of China.
- Cite (Informal):
- CCL25-Eval任务四系统报告:基于RAG与谓词相似性方法的叙实性检测智能体 (Wang et al., CCL 2025)
- Copy Citation:
- PDF:
- https://aclanthology.org/2025.ccl-2.18.pdf