README.md

---
license: apache-2.0
language:
- en
- zh
---
# GraphGen-Data

<!-- Provide a quick summary of the dataset. -->

## Data Description

GraphGen-Data is the dataset for verification in the paper "[GraphGen: Enhancing Supervised Fine-Tuning for LLMs with Knowledge-Driven Synthetic Data Generation](https://arxiv.org/abs/2505.20416)".
It involves three domains:
- Agricultural(SeedEval)
- Medical(PQArefEval)
- General(HotpotEval)

GraphGen is a framework for synthetic data generation guided by knowledge graphs. We released our code in [Github](https://github.com/open-sciencelab/GraphGen).

## Source Data

<!-- This section describes the source data (e.g. news text and headlines, social media posts, translated sentences, ...). -->

SeedEval is adapted from [SeedBench](https://arxiv.org/abs/2505.13220), a benchmark with 11 tasks related to seed knowledge. For this study, we selected Task QA–4 (covering one-shot and zero-shot scenarios) related to textual knowledge question answering.
PQArefEval is derived from [PQAref](https://arxiv.org/abs/2407.05015), from which we extracted 5,818 instances for our analysis.
[HotpotQA](https://arxiv.org/abs/1809.09600) is a dataset for diverse, explainable multi-hop question answering, where questions require integrating information from multiple sources. We used the test set of HotpotQA as the new evaluation dataset, HotpotEval. 
Each dataset comprises two components: the QA test set and the corresponding source texts.
The Corpus for SeedEval is provided by anonymous agricultural experts and cannot be made public due to confidentiality restrictions.
The Corpus for PQArefEval and HotpotEval are constructed from the original references of PQAref and HotpotQA, respectively. 

## Citation

<!-- If there is a paper or blog post introducing the dataset, the APA and Bibtex information for that should go in this section. -->

**BibTeX:**

```
@misc{chen2025graphgenenhancingsupervisedfinetuning,
      title={GraphGen: Enhancing Supervised Fine-Tuning for LLMs with Knowledge-Driven Synthetic Data Generation}, 
      author={Zihong Chen and Wanli Jiang and Jinzhe Li and Zhonghang Yuan and Huanjun Kong and Wanli Ouyang and Nanqing Dong},
      year={2025},
      eprint={2505.20416},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2505.20416}, 
}
```