File size: 3,175 Bytes
641d807
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
77d83da
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
---
dataset_info:
  features:
  - name: images
    sequence: image
  - name: question
    dtype: string
  - name: answers
    sequence: string
  - name: correct_answer
    dtype: string
  - name: question_type
    dtype: string
  splits:
  - name: train
    num_bytes: 5167070090.512
    num_examples: 172384
  - name: static
    num_bytes: 3140831722.665
    num_examples: 127405
  - name: val
    num_bytes: 305661617.158
    num_examples: 4001
  - name: test
    num_bytes: 125653489.0
    num_examples: 150
  download_size: 2182325666
  dataset_size: 8739216919.335
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: static
    path: data/static-*
  - split: val
    path: data/val-*
  - split: test
    path: data/test-*
---

# SAT-v2 Dataset

## Paper

**SAT: Dynamic Spatial Aptitude Training for Multimodal Language Models**

This dataset is part of the SAT (Spatial Aptitude Training) project, which introduces a dynamic benchmark for evaluating and improving spatial reasoning capabilities in multimodal language models.

- **Project Page**: [https://arijitray.com/SAT/](https://arijitray.com/SAT/)
- **Paper**: [arXiv:2412.07755](https://arxiv.org/abs/2412.07755)

## Dataset Description

SAT-v2 is a comprehensive spatial reasoning benchmark containing over 300,000 questions across multiple splits. The dataset tests various aspects of spatial understanding including perspective-taking, object relationships, and dynamic scene understanding.

## Loading the Dataset

```python
from datasets import load_dataset

# Load the training split
dataset = load_dataset("array/SAT-v2", split="train")

# Or load a specific split
val_dataset = load_dataset("array/SAT-v2", split="val")
static_dataset = load_dataset("array/SAT-v2", split="static")
test_dataset = load_dataset("array/SAT-v2", split="test")

# Access a sample
sample = dataset[0]
print(sample["question"])
print(sample["answers"])
print(sample["correct_answer"])
```

## Dataset Splits

- **train**: 172,384 examples - Dynamic training questions
- **static**: 127,405 examples - Static spatial reasoning questions
- **val**: 4,001 examples - Validation set
- **test**: 150 examples - Test set

**Important Note on Test Set Evaluation:** When evaluating on the test set, please use circular evaluation by switching the position of the correct answer to avoid position bias. If you're using lmms-eval, refer to the implementation here: [https://github.com/arijitray1993/lmms-eval/tree/main/lmms_eval/tasks/sat_real](https://github.com/arijitray1993/lmms-eval/tree/main/lmms_eval/tasks/sat_real)

## Citation

If you use this dataset, please cite:

```bibtex
@misc{ray2025satdynamicspatialaptitude,
      title={SAT: Dynamic Spatial Aptitude Training for Multimodal Language Models},
      author={Arijit Ray and Jiafei Duan and Ellis Brown and Reuben Tan and Dina Bashkirova and Rose Hendrix and Kiana Ehsani and Aniruddha Kembhavi and Bryan A. Plummer and Ranjay Krishna and Kuo-Hao Zeng and Kate Saenko},
      year={2025},
      eprint={2412.07755},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2412.07755},
}
```