README.md

---
license: cc-by-4.0
language:
- en
tags:
- pubmed
- embeddings
- medcpt
- biomedical
- retrieval
- rag
- medical
pretty_name: PubMedAbstractsSubsetEmbedded
---

# PubMed Abstracts Subset with MedCPT Embeddings (float32)

This dataset contains a probabilistic sample of ~2.4 million PubMed abstracts, enriched with precomputed dense embeddings (title + abstract), from the **`ncbi/MedCPT-Article-Encoder`** model. It is derived from public metadata made available via the [National Library of Medicine (NLM)](https://pubmed.ncbi.nlm.nih.gov/) and was used in the paper [*Efficient and Reproducible Biomedical QA using Retrieval-Augmented Generation*](https://arxiv.org/abs/2505.07917).

Each entry includes:
- `title`: Title of the publication
- `abstract`: Abstract content
- `PMID`: PubMed identifier
- `embedding`: 768-dimensional float32 vector from MedCPT

---

## 🔍 How to Access

### ▶️ Option 1: Load via Hugging Face `datasets`

```python
from datasets import load_dataset

dataset = load_dataset("slinusc/PubMedAbstractsSubsetEmbedded", streaming=True)

for doc in dataset:
    print(doc["PMID"], doc["embedding"][:5])  # print first 5 dims
    break
```

> Each vector is stored as a list of 768 `float32` values (compact, no line breaks).

---

### 💾 Option 2: Git Clone with Git LFS

```bash
git lfs install
git clone https://huggingface.co/datasets/slinusc/PubMedAbstractsSubsetEmbedded
cd PubMedAbstractsSubsetEmbedded
```

---

## 📦 Format

Each file is a `.jsonl` (JSON Lines) file, where each line is a valid JSON object:

```json
{
  "title": "...",
  "abstract": "...",
  "PMID": 36464820,
  "embedding": [-0.1952481, ... , 0.2887376]
}
```

> The embeddings are 768-dimensional dense vectors, serialized as 32-bit floats.

---

## 📚 Source and Licensing

This dataset is derived from public domain PubMed metadata (titles and abstracts), redistributed in accordance with [NLM data usage policies](https://www.nlm.nih.gov/databases/download/data_distrib_main.html).  
MedCPT embeddings were generated using the [ncbi/MedCPT-Article-Encoder](https://huggingface.co/ncbi/MedCPT-Article-Encoder) model.

---

## 📣 Citation

If you use this dataset or the included MedCPT embeddings, please cite:

> **Stuhlmann et al. (2025)**  
> *Efficient and Reproducible Biomedical Question Answering using Retrieval Augmented Generation*  
> [arXiv:2505.07917](https://arxiv.org/abs/2505.07917)  
> [https://github.com/slinusc/medical_RAG_system](https://github.com/slinusc/medical_RAG_system)

---

## 🏷️ Version

- `v1.0` – Initial release (2.39M samples, 24 JSONL files, float32 embeddings, ~23 GB total)

---

## 📬 Contact

Maintained by [@slinusc](https://huggingface.co/slinusc).  
For questions or collaborations, open a discussion on the HF Hub.