# Graphgen Eval

> Evaluates the ability of LLMs to answer knowledge-intensive questions across atomic, aggregated, and multi-hop reasoning scenarios in agricultural, medical, and general domains. It measures how well supervised fine-tuning with synthetic knowledge-graph data improves closed-book QA performance. Use when the user wants to benchmark on SeedEval, PQArefEval, HotpotEval, or asks about evaluating this task. Reports ROUGE-F.

- Skill: `qhjqhj00/graphgen-eval` (Agent Skill)
- Install (CLI): `npx skillmds add qhjqhj00/graphgen-eval`
- Raw SKILL.md: https://api.skillmd.com/api/skills/qhjqhj00/graphgen-eval/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Productivity
- Author: qhjqhj00 (https://skillmd.com/u/qhjqhj00)
- Updated: 2026-09-08
- Page: https://skillmd.com/skills/qhjqhj00/graphgen-eval

---


# graphgen-eval

> GraphGen: Enhancing Supervised Fine-Tuning for LLMs with Knowledge-Driven Synthetic Data Generation — Chen et al. (2025) (arXiv:2505.20416, 2025)

## What this evaluates

Evaluates the ability of LLMs to answer knowledge-intensive questions across atomic, aggregated, and multi-hop reasoning scenarios in agricultural, medical, and general domains. It measures how well supervised fine-tuning with synthetic knowledge-graph data improves closed-book QA performance.

## Datasets

- **SeedEval** — total 582; splits: test (582)
- **PQArefEval** — total 5815; splits: test (5815)
- **HotpotEval** — total 7405; splits: test (7405)

## Metrics

- `ROUGE-F` **(primary)** — range: [0, 1]
  - F1 score measuring the overlap of n-grams between the generated answer and the ground truth reference.

## Input / output format

**Input**: A natural language question for closed-book QA.

**Output**: A natural language text answer.

## Scoring recipe

```python
def evaluate(predictions, references):
    scores = []
    for pred, ref in zip(predictions, references):
        score = compute_rouge_f1(pred, ref)  # ROUGE-F as specified in the paper
        scores.append(score)
    return sum(scores) / len(scores)
```

## Common pitfalls

- The evaluation is strictly closed-book; allowing external retrieval during inference invalidates the benchmark.
- ROUGE-F measures lexical overlap, not factual correctness or reasoning depth, so high scores do not guarantee accurate multi-hop answers.
- Baseline methods can cause performance degradation on aggregated QA tasks due to single-segment generation limits, which should be attributed to data synthesis flaws rather than the evaluation metric itself.

## Evidence (verbatim from paper)

> We use ROUGE-F as the metric. The baseline methods exhibit varying performance across the three datasets, while GraphGen consistently achieves optimal results.

## Citation

```bibtex
@misc{chen2025graphgen,
  title={GraphGen: Enhancing Supervised Fine-Tuning for LLMs with Knowledge-Driven Synthetic Data Generation},
  author={Chen et al. (2025)},
  year={2025},
  note={arXiv:2505.20416}
}
```

- arXiv: 2505.20416