Can Frontier LLMs Replace Annotators in Biomedical Text Mining? Analyzing Challenges and Exploring Solutions
Yichong Zhao, Susumu Goto · 2025
Read Can Frontier LLMs Replace Annotators in Biomedical Text Mining? Analyzing Challenges and Exploring Solutions by Yichong Zhao, Susumu Goto (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-2503-03261.txt.
Description
Yichong Zhao and Susumu Goto ask whether top language models can take over biomedical text tagging in a 2025 paper. Past tests showed the models lagging, and the authors dig into why.
Three failure roots stand out: models miss quiet data set habits, stiff output formats cramp their reasoning, and they drift from tagging rulebooks. Prompt fixes aimed at each root, plus a pipeline that pulls rules straight from the guide, lift the models to par with fine tuned specialists, with little hand tagged data. A small model trained only on model made tags also reached usable quality.
Catalog record
| CREATOR | Yichong Zhao, Susumu Goto |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 50 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-03261