Benchmark
Massive-STEPS Benchmark
What is Massive-STEPS?
Massive-STEPS is a large-scale semantic trajectories dataset designed for understanding and predicting Point-of-Interest (POI) check-ins.
- Released
- 2025-05-16
- Modality
- multimodal
- Evaluates
- 多模态, Multimodal, 推理, Reasoning, VQA, Embodied Decision Making, 物理智能, Embodied AI, 逻辑推理, 具身交互, Embodied Interaction, 跨模态推理, Cross-modal Reasoning, 不支持, Unsupported
- Openness
- open
- Importer
- OpenCompass Hub
- Reported scores
- 0
Source provenance
- Original evidence https://hub.opencompass.org.cn/dataset-detail/Massive-STEPS
Massive-STEPS paper, code and dataset
- Paper https://arxiv.org/pdf/2505.11239
- Code https://github.com/cruiseresearchgroup/Massive-STEPS
- Dataset https://huggingface.co/datasets/CRUISEResearchGroup/Massive-STEPS-Beijing
- Dataset https://huggingface.co/datasets/CRUISEResearchGroup/Massive-STEPS-Istanbul
- Dataset https://huggingface.co/datasets/CRUISEResearchGroup/Massive-STEPS-Jakarta
- Dataset https://huggingface.co/datasets/CRUISEResearchGroup/Massive-STEPS-Kuwait-City
- Dataset https://huggingface.co/datasets/CRUISEResearchGroup/Massive-STEPS-Melbourne
- Dataset https://huggingface.co/datasets/CRUISEResearchGroup/Massive-STEPS-Moscow
- Dataset https://huggingface.co/datasets/CRUISEResearchGroup/Massive-STEPS-New-York
- Dataset https://huggingface.co/datasets/CRUISEResearchGroup/Massive-STEPS-Petaling-Jaya
- Dataset https://huggingface.co/datasets/CRUISEResearchGroup/Massive-STEPS-Sao-Paulo
- Dataset https://huggingface.co/datasets/CRUISEResearchGroup/Massive-STEPS-Shanghai
- Dataset https://huggingface.co/datasets/CRUISEResearchGroup/Massive-STEPS-Sydney
- Dataset https://huggingface.co/datasets/CRUISEResearchGroup/Massive-STEPS-Tokyo
No reported scores are on record for this benchmark yet.