ik-ram28/

synthetic-NER-dataset

	Synthetic Biomedical NER Dataset




	
	Dataset Description

This dataset contains synthetically generated biomedical text annotations for Named Entity Recognition (NER), created to augment a limited training corpus for the EvalLLM 2025 challenge. The synthetic data was generated using GPT-4.1 to enhance the original 40-document training set.

Original training documents: 40 annotated documents Synthetic documents generated: 1,748 documents Generation… See the full description on the dataset page: https://huggingface.co/datasets/ik-ram28/synthetic-NER-dataset.