odl-raiser/Geoint

---
license: apache-2.0
task_categories:
- question-answering
tags:
- math
- geometry
- visual-question-answering
---

### Dataset Summary

**Geoint** is a comprehensive benchmark dataset explicitly designed for formal geometric problemsolving. Geoint encompasses 1,885 carefully curated geometric questions across diverse categories including plane, spatial, and solid geometry problems. Each problem is richly annotated with both structured textual descriptions and accompanying visual diagrams to support multimodal understanding. Furthermore, Geoint leverages the Lean 4 proof assistant to formally represent geometric elements and relationships, enabling rigorous and complete formal reasoning within a verifiable framework.

You can load the dataset using:

```python

from datasets import load_dataset

dataset = load_dataset("your-username/Geoint", split="train")

```

### Data Fields

| Field | Type | Description |
|------------------|-------------|-------------|
| `id` | `string` | Unique identifier for each problem |
| `question` | `string` | Text description of the geometric problem |
| `response` | `string` | Solution text and its corresponding Lean 4 code|
| `question_type` | `string` | Problem type: *Free* or *Proof* |
| `assistant_line` | `bool` | Whether auxiliary lines are needed for solving |
| `question_image` | `binary` | Problem diagram image (PNG format) |
| `assistant_image`| `binary` or `None` | Image with auxiliary lines (PNG format) |

### Citation Information

Please consider citing [our paper](https://arxiv.org/pdf/2508.03173) if you find this dataset useful:

```
@misc{wei2025geointr1formalizingmultimodalgeometric,
title={Geoint-R1: Formalizing Multimodal Geometric Reasoning with Dynamic Auxiliary Constructions},
author={Jingxuan Wei and Caijun Jia and Qi Chen and Honghao He and Linzhuang Sun and Conghui He and Lijun Wu and Bihui Yu and Cheng Tan},
year={2025},
eprint={2508.03173},
archivePrefix={arXiv},
primaryClass={cs.AI},
url={https://arxiv.org/abs/2508.03173},
}
```