跳到正文
arXiv cs.CL· Marc Brinner, Philipp Cimiano, Karolina Hanna Czok, Regina Stodden, Viju Sudhi, Sina Zarrie{\ss}·· 2 天前AI 评分18

NER-D:面向德语 NER 的跨领域基准

NER-D: A cross-domain benchmark for German NER

AI 导读

NER-D 是一个覆盖 9 个领域、13 个数据集的德语 NER 跨领域基准,涵盖通用/政治、维基百科、法律、医疗、历史、金融、农业与考古。在无优化设置下,GPT-5.4 mini 在 13 个数据集中的 6 个上优于小型开源模型,但 LoRA 微调后小型模型可显著超越该专有基线。研究还分析了领域与各任务标签集对不同评测设置下性能的影响。

正文 · 原文

View PDF HTML (experimental)

Abstract:We present NER-D, a cross-domain benchmark for German NER comprising 13 datasets spanning 9 domains, including General/Politics, Wikipedia, Legal, Medical, Historical, Financial, Agriculture and Archaeology. We examine the performance of both proprietary LLMs and state-of-the-art small open-source models using normal prompting and further examine the performance gains achievable via in-context learning and LoRA fine-tuning. Our results show that without any sort of optimization a proprietary model (GPT-5.4 mini) outperforms the smaller open-source models on 6 out of 13 datasets, but LoRA fine-tuning enables the smaller models to substantially outperform the proprietary baseline. We further analyze the influence of the domain and the tagset used in each task on the performance achieved in the different evaluation settings. Our benchmark will be available as a resource for the development of future German NER methods on both general and specialized domains.\footnote{We make our code, data, and models publicly available upon acceptance of this paper.
Subjects: Computation and Language (cs.CL)
Cite as: arXiv:2611.18257 [cs.CL]
  (or arXiv:2611.18257v1 [cs.CL] for this version)
  https://doi.org/10.48550/arXiv.2611.18257

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Marc Felix Brinner [view email]
[v1] Fri, 13 Nov 2026 14:54:21 UTC (48 KB)

来源:arXiv cs.CL · arxiv.org