Clinical named entity recognition from dental progress notes is challenging because documentation is highly unstructured, domain-specific, and often privacy-sensitive. We developed a locally deployable framework that enables small language models to self-generate, verify, refine, and evaluate entity-specific prompts for extracting multiple clinical entities from dental notes. Using 1,200 annotated notes, we evaluated candidate open-weight models with multi-prompt ensemble inference and further adapted selected models using QLoRA-based supervised fine-tuning and direct preference optimization. Model performance varied substantially, highlighting the need for task-specific evaluation rather than reliance on generic benchmarks. Qwen2.5-14B-Instruct achieved the strongest baseline performance. After DPO, Qwen2.5-14B-Instruct and Llama-3.1-8B-Instruct achieved micro/macro F1 scores of 0.864/0.837 and 0.806/0.797, respectively. These findings suggest that automated prompt optimization combined with lightweight preference-based post-training can support scalable clinical information extraction using locally deployed small language models.
@article{arxiv.2605.04221,
title = {Self-Prompting Small Language Models for Privacy-Sensitive Clinical Information Extraction},
author = {Yao-Shun Chuang and Tushti Mody and Uday Pratap Singh and Shirindokht Shiraz and Chun-Teh Lee and Ryan Brandon and Muhammad F Walji and Xiaoqian Jiang and Bunmi Tokede},
journal= {arXiv preprint arXiv:2605.04221},
year = {2026}
}