Datasets:
| license: mit | |
| task_categories: | |
| - text-generation | |
| language: | |
| - en | |
| tags: | |
| - medical | |
| - biomedical | |
| - abstract | |
| - conclusion-generation | |
| pretty_name: MedConclusion | |
| size_categories: | |
| - 1M<n<10M | |
| # MedConclusion | |
| **MedConclusion** is a large-scale dataset of 5.7M PubMed structured abstracts for biomedical conclusion generation. Each instance pairs the non-conclusion sections of an abstract with the original author-written conclusion, providing naturally occurring supervision for evidence-to-conclusion reasoning. MedConclusion also includes journal-level metadata such as biomedical category and SJR, enabling subgroup analysis across biomedical domains. | |
| This repository contains the full version of the dataset. For faster evaluation and model prototyping, please check out the [**Compact Version**](https://huggingface.co/datasets/harvardairobotics/MedConclusion-Compact). | |
| - **Train**: 70% | |
| - **Validation**: 10% | |
| - **Test**: 20% | |
| ## Benchmark Information | |
| This dataset is introduced in the paper [MedConclusion: A Benchmark for Biomedical Conclusion Generation from Structured Abstracts](https://arxiv.org/abs/2604.06505). | |
| ## Citation | |
| ```bibtex | |
| @article{li2026medconclusion, | |
| title={MedConclusion: A Benchmark for Biomedical Conclusion Generation from Structured Abstracts}, | |
| author={Li, Weiyue and Qian, Ruizhi and Li, Yi and Li, Yongce and Long, Yunfan and Cai, Jiahui and Luo, Yan and Wang, Mengyu}, | |
| journal={arXiv preprint arXiv:2604.06505}, | |
| year={2026} | |
| } | |
| ``` | |