DHd 2025: Workshop eScriptorium meets LLMs

Hands-On 3: LLMs for Named Entity Recognition on OCR output

This experimental repository and jupyter notebook demonstrates how to apply large language models (LLMs) to Named Entity Recognition (NER) tasks.

It is based on these papers:

• Zhang, R., Li, Y., Ma, Y., Zhou, M., & Zou, L. (2023). LLMaAA: Making Large Language Models as Active Annotators. In H. Bouamor, J. Pino, & K. Bali (Hrsg.), Findings of the Association for Computational Linguistics: EMNLP 2023 (S. 13088–13103). Association for Computational Linguistics. https://doi.org/10.18653/v1/2023.findings-emnlp.872
  • GitHub repo: https://github.com/ridiculouz/LLMaAA/
• Dalfsen, A. van, Karsdorp, F., Bagheri, A., Engelen, T. van, Mentink, D., & Stronks, E. (2024). Direct and Indirect Annotation with Generative AI: A Case Study into Finding Animals and Plants in Historical Text. Proceedings of the Computational Humanities Research Conference, 2024. https://ceur-ws.org/Vol-3834/paper74.pdf
  • GitHub repo: https://github.com/trister95/direct-and-indirect-annotation

Repository structure

project/
├── README.md
├── LLMs4NER/
│   ├── DHd_2025_LLMs4NER.ipynb
│   └── data/
│       ├── labeled_data.json
└── LICENSE