| language: | |
| - en | |
| license: cc-by-4.0 | |
| size_categories: | |
| - n<1K | |
| pretty_name: Rocket Science | |
| configs: | |
| - config_name: default | |
| data_files: | |
| - split: train | |
| path: data/train-* | |
| task_categories: | |
| - image-text-to-text | |
| tags: | |
| - spatial-reasoning | |
| - vlm | |
| - benchmark | |
| dataset_info: | |
| features: | |
| - name: text1 | |
| dtype: string | |
| - name: text2 | |
| dtype: string | |
| - name: Label | |
| dtype: string | |
| - name: Category | |
| dtype: string | |
| - name: image1 | |
| dtype: image | |
| - name: image2 | |
| dtype: image | |
| splits: | |
| - name: train | |
| num_bytes: 772658832 | |
| num_examples: 241 | |
| download_size: 771816767 | |
| dataset_size: 772658832 | |
| # Dataset Card for Rocket Science | |
| Rocket Science is a benchmark that tests for understanding of spatial relations in Vision Language Models. | |
| # Dataset Description | |
| We propose RocketScience, an open-source contrastive VLM benchmark that tests for spatial relation understanding. It is comprised of entirely new real-world image-text pairs covering mostly relative spatial understanding and the order of objects. The benchmark is designed to be very easy for humans and hard for the current generation of VLMs, and this is empirically verified. Our results show a striking lack of spatial relation understanding in open source and frontier commercial VLMs and a surprisingly high performance of reasoning models. Additionally, we perform a disentanglement analysis to separate the contributions of object localization and spatial reasoning in chain-of-thought-based models and find that the performance on the benchmark is bottlenecked by spatial reasoning and not object localization capabilities. | |
| - **Curated by:** Nils Hoehing | |
| - **Language(s) (NLP):** English | |
| - **License:** CC-BY-4.0 | |
| # Dataset Sources | |
| - **Repository:** https://github.com/nilshoehing/rocketscience | |
| - **Paper:** https://huggingface.co/papers/2509.02175 | |
| # Sample Usage | |
| The dataset can be loaded using the Hugging Face `datasets` library: | |
| ```python | |
| from datasets import load_dataset | |
| dataset = load_dataset("nilshoehing/rocketsciencebench") | |
| ``` | |
| # Citation | |
| Please cite the following paper if you use this dataset: | |
| ```bibtex | |
| @article{hoehing2025understanding, | |
| title={Understanding Space Is Rocket Science - Only Top Reasoning Models Can Solve Spatial Understanding Tasks}, | |
| author={Hoehing, Nils and Likhitha, D and Woehler, Thomas and Beham, Michael and Kuehne, Axel}, | |
| journal={arXiv preprint arXiv:2509.02175}, | |
| year={2025}, | |
| url={https://huggingface.co/papers/2509.02175} | |
| } | |
| ``` |