Retrieval augmented generation based dynamic prompting for few-shot biomedical named entity recognition using large language models
Yao Ge, Sudeshna Das, Yuting Guo, Abeed Sarker · 2025
Read Retrieval augmented generation based dynamic prompting for few-shot biomedical named entity recognition using large language models by Yao Ge, Sudeshna Das, Yuting Guo, Abeed Sarker (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2508-06504.txt.
Description
Tagging diseases and drugs in medical text is essential drudge work, and language models do it unevenly when examples are few. Yao Ge and co-authors fix the prompt instead of the model in a 2025 paper: for each new text, retrieve the most similar labeled examples and rebuild the prompt on the fly.
Across five biomedical data sets, a well-structured fixed prompt already lifted scores by about twelve percent for a leading model. Dynamic prompts added more, with two retrieval styles leading the pack in five- and ten-example settings. Matching examples to the text at hand, rather than freezing a generic set, proved the difference. The approach needs no retraining, which suits fast-moving clinical vocabularies.
Catalog record
| CREATOR | Yao Ge, Sudeshna Das, Yuting Guo, Abeed Sarker |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 84 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2508-06504