README.md

---
license: cc-by-4.0
language:
- zh
task_categories:
- question-answering
- text-generation
task_ids:
- multiple-choice-qa
pretty_name: NursData-MCQ
size_categories:
- 1K<n<10K
---

# NursData-MCQ

## Dataset Description

NursData-MCQ is a Chinese nursing multiple-choice benchmark for evaluating large language models in fundamental nursing knowledge and nursing-domain reasoning. It was used as the automated evaluation dataset in the EviNurse study.

EviNurse is a domain-specific large language model for evidence-based nursing, developed on Qwen3-32B with supervised fine-tuning and retrieval-augmented generation. In the manuscript, automated evaluation was conducted with **3,438** multiple-choice questions to assess the model's basic nursing capability and compare it with general-purpose LLMs.

Project code is available at:

https://github.com/Creeper12345/EviNurse

Model:

https://huggingface.co/Agnania/EviNurse-32B

## Dataset File

| File | Split | Questions | Description |
| --- | --- | ---: | --- |
| `evinurse_automated_eval_3438.json` | test | 3,438 | Chinese nursing multiple-choice benchmark for automated evaluation. |

## Data Fields

Each record contains:

```json
{
  "id": "question identifier",
  "question": "question stem",
  "options": {
    "A": "option A",
    "B": "option B",
    "C": "option C",
    "D": "option D",
    "E": "option E"
  },
  "answer": "standard answer"
}
```

## Download

Download the complete dataset with:

```python
from huggingface_hub import snapshot_download

snapshot_download(
    repo_id="Agnania/NursData-MCQ",
    repo_type="dataset"
)
```

Alternatively, using the Hugging Face CLI:

```bash
hf download Agnania/NursData-MCQ \
  --repo-type dataset
```

You can also load the JSON file directly:

```python
import json
from pathlib import Path

path = Path("evinurse_automated_eval_3438.json")
with path.open(encoding="utf-8") as f:
    data = json.load(f)

print(len(data))  # 3438
print(data[0])
```

## Evaluation

For automated multiple-choice evaluation, a model should output one final option letter. Accuracy is calculated by exact match against the `answer` field.

## License

This dataset is released under the Creative Commons Attribution 4.0 International License (CC BY 4.0).