Benchmarking Spatiotemporal Reasoning in LLMs and Reasoning Models: Capabilities and Challenges
Paper • 2505.11618 • Published
Datasets:
Tasks:
Question Answering
Modalities:
Text
Formats:
csv
Languages:
English
Size:
10K - 100K
ArXiv:
License:
Unnamed: 0 int64 0 0 | query stringlengths 639 5.27k | answer stringlengths 5 414 |
|---|---|---|
0 |
Help me answer question regarding spatial relationship in a 2D plane:
Given Information:
You will receive a series of object trajectory and the corresponding timestamps of the coordinates in the trajectory. You can treat the trajectory as linestring.
Sensor A: [(x_1, y_1), (x_2, y_2), ..., (x_n, y_n)]
... | [1.0] |
0 |
Help me answer question regarding spatial relationship in a 2D plane:
Given Information:
You will receive a series of object trajectory and the corresponding timestamps of the coordinates in the trajectory. You can treat the trajectory as linestring.
Sensor A: [(x_1, y_1), (x_2, y_2), ..., (x_n, y_n)]
... | [1.0] |
0 |
Help me answer question regarding spatial relationship in a 2D plane:
Given Information:
You will receive a series of object trajectory and the corresponding timestamps of the coordinates in the trajectory. You can treat the trajectory as linestring.
Sensor A: [(x_1, y_1), (x_2, y_2), ..., (x_n, y_n)]
... | [1.0] |
0 |
Help me answer question regarding spatial relationship in a 2D plane:
Given Information:
You will receive a series of object trajectory and the corresponding timestamps of the coordinates in the trajectory. You can treat the trajectory as linestring.
Sensor A: [(x_1, y_1), (x_2, y_2), ..., (x_n, y_n)]
... | [0.0] |
0 |
Help me answer question regarding spatial relationship in a 2D plane:
Given Information:
You will receive a series of object trajectory and the corresponding timestamps of the coordinates in the trajectory. You can treat the trajectory as linestring.
Sensor A: [(x_1, y_1), (x_2, y_2), ..., (x_n, y_n)]
... | [1.0] |
0 |
Help me answer question regarding spatial relationship in a 2D plane:
Given Information:
You will receive a series of object trajectory and the corresponding timestamps of the coordinates in the trajectory. You can treat the trajectory as linestring.
Sensor A: [(x_1, y_1), (x_2, y_2), ..., (x_n, y_n)]
... | [0.0] |
0 |
Help me answer question regarding spatial relationship in a 2D plane:
Given Information:
You will receive a series of object trajectory and the corresponding timestamps of the coordinates in the trajectory. You can treat the trajectory as linestring.
Sensor A: [(x_1, y_1), (x_2, y_2), ..., (x_n, y_n)]
... | [1.0] |
0 |
Help me answer question regarding spatial relationship in a 2D plane:
Given Information:
You will receive a series of object trajectory and the corresponding timestamps of the coordinates in the trajectory. You can treat the trajectory as linestring.
Sensor A: [(x_1, y_1), (x_2, y_2), ..., (x_n, y_n)]
... | [0.0] |
0 |
Help me answer question regarding spatial relationship in a 2D plane:
Given Information:
You will receive a series of object trajectory and the corresponding timestamps of the coordinates in the trajectory. You can treat the trajectory as linestring.
Sensor A: [(x_1, y_1), (x_2, y_2), ..., (x_n, y_n)]
... | [1.0] |
0 |
Help me answer question regarding spatial relationship in a 2D plane:
Given Information:
You will receive a series of object trajectory and the corresponding timestamps of the coordinates in the trajectory. You can treat the trajectory as linestring.
Sensor A: [(x_1, y_1), (x_2, y_2), ..., (x_n, y_n)]
... | [1.0] |
0 |
Help me answer question regarding spatial relationship in a 2D plane:
Given Information:
You will receive a series of object trajectory and the corresponding timestamps of the coordinates in the trajectory. You can treat the trajectory as linestring.
Sensor A: [(x_1, y_1), (x_2, y_2), ..., (x_n, y_n)]
... | [1.0] |
0 |
Help me answer question regarding spatial relationship in a 2D plane:
Given Information:
You will receive a series of object trajectory and the corresponding timestamps of the coordinates in the trajectory. You can treat the trajectory as linestring.
Sensor A: [(x_1, y_1), (x_2, y_2), ..., (x_n, y_n)]
... | [1.0] |
0 |
Help me answer question regarding spatial relationship in a 2D plane:
Given Information:
You will receive a series of object trajectory and the corresponding timestamps of the coordinates in the trajectory. You can treat the trajectory as linestring.
Sensor A: [(x_1, y_1), (x_2, y_2), ..., (x_n, y_n)]
... | [1.0] |
0 |
Help me answer question regarding spatial relationship in a 2D plane:
Given Information:
You will receive a series of object trajectory and the corresponding timestamps of the coordinates in the trajectory. You can treat the trajectory as linestring.
Sensor A: [(x_1, y_1), (x_2, y_2), ..., (x_n, y_n)]
... | [1.0] |
0 |
Help me answer question regarding spatial relationship in a 2D plane:
Given Information:
You will receive a series of object trajectory and the corresponding timestamps of the coordinates in the trajectory. You can treat the trajectory as linestring.
Sensor A: [(x_1, y_1), (x_2, y_2), ..., (x_n, y_n)]
... | [0.0] |
0 |
Help me answer question regarding spatial relationship in a 2D plane:
Given Information:
You will receive a series of object trajectory and the corresponding timestamps of the coordinates in the trajectory. You can treat the trajectory as linestring.
Sensor A: [(x_1, y_1), (x_2, y_2), ..., (x_n, y_n)]
... | [0.0] |
0 |
Help me answer question regarding spatial relationship in a 2D plane:
Given Information:
You will receive a series of object trajectory and the corresponding timestamps of the coordinates in the trajectory. You can treat the trajectory as linestring.
Sensor A: [(x_1, y_1), (x_2, y_2), ..., (x_n, y_n)]
... | [0.0] |
0 |
Help me answer question regarding spatial relationship in a 2D plane:
Given Information:
You will receive a series of object trajectory and the corresponding timestamps of the coordinates in the trajectory. You can treat the trajectory as linestring.
Sensor A: [(x_1, y_1), (x_2, y_2), ..., (x_n, y_n)]
... | [1.0] |
0 |
Help me answer question regarding spatial relationship in a 2D plane:
Given Information:
You will receive a series of object trajectory and the corresponding timestamps of the coordinates in the trajectory. You can treat the trajectory as linestring.
Sensor A: [(x_1, y_1), (x_2, y_2), ..., (x_n, y_n)]
... | [0.0] |
0 |
Help me answer question regarding spatial relationship in a 2D plane:
Given Information:
You will receive a series of object trajectory and the corresponding timestamps of the coordinates in the trajectory. You can treat the trajectory as linestring.
Sensor A: [(x_1, y_1), (x_2, y_2), ..., (x_n, y_n)]
... | [1.0] |
0 |
Help me answer question regarding spatial relationship in a 2D plane:
Given Information:
You will receive a series of object trajectory and the corresponding timestamps of the coordinates in the trajectory. You can treat the trajectory as linestring.
Sensor A: [(x_1, y_1), (x_2, y_2), ..., (x_n, y_n)]
... | [1.0] |
0 |
Help me answer question regarding spatial relationship in a 2D plane:
Given Information:
You will receive a series of object trajectory and the corresponding timestamps of the coordinates in the trajectory. You can treat the trajectory as linestring.
Sensor A: [(x_1, y_1), (x_2, y_2), ..., (x_n, y_n)]
... | [0.0] |
0 |
Help me answer question regarding spatial relationship in a 2D plane:
Given Information:
You will receive a series of object trajectory and the corresponding timestamps of the coordinates in the trajectory. You can treat the trajectory as linestring.
Sensor A: [(x_1, y_1), (x_2, y_2), ..., (x_n, y_n)]
... | [0.0] |
0 |
Help me answer question regarding spatial relationship in a 2D plane:
Given Information:
You will receive a series of object trajectory and the corresponding timestamps of the coordinates in the trajectory. You can treat the trajectory as linestring.
Sensor A: [(x_1, y_1), (x_2, y_2), ..., (x_n, y_n)]
... | [0.0] |
0 |
Help me answer question regarding spatial relationship in a 2D plane:
Given Information:
You will receive a series of object trajectory and the corresponding timestamps of the coordinates in the trajectory. You can treat the trajectory as linestring.
Sensor A: [(x_1, y_1), (x_2, y_2), ..., (x_n, y_n)]
... | [1.0] |
0 |
Help me answer question regarding spatial relationship in a 2D plane:
Given Information:
You will receive a series of object trajectory and the corresponding timestamps of the coordinates in the trajectory. You can treat the trajectory as linestring.
Sensor A: [(x_1, y_1), (x_2, y_2), ..., (x_n, y_n)]
... | [0.0] |
0 |
Help me answer question regarding spatial relationship in a 2D plane:
Given Information:
You will receive a series of object trajectory and the corresponding timestamps of the coordinates in the trajectory. You can treat the trajectory as linestring.
Sensor A: [(x_1, y_1), (x_2, y_2), ..., (x_n, y_n)]
... | [1.0] |
0 |
Help me answer question regarding spatial relationship in a 2D plane:
Given Information:
You will receive a series of object trajectory and the corresponding timestamps of the coordinates in the trajectory. You can treat the trajectory as linestring.
Sensor A: [(x_1, y_1), (x_2, y_2), ..., (x_n, y_n)]
... | [1.0] |
0 |
Help me answer question regarding spatial relationship in a 2D plane:
Given Information:
You will receive a series of object trajectory and the corresponding timestamps of the coordinates in the trajectory. You can treat the trajectory as linestring.
Sensor A: [(x_1, y_1), (x_2, y_2), ..., (x_n, y_n)]
... | [1.0] |
0 |
Help me answer question regarding spatial relationship in a 2D plane:
Given Information:
You will receive a series of object trajectory and the corresponding timestamps of the coordinates in the trajectory. You can treat the trajectory as linestring.
Sensor A: [(x_1, y_1), (x_2, y_2), ..., (x_n, y_n)]
... | [0.0] |
0 |
Help me answer question regarding spatial relationship in a 2D plane:
Given Information:
You will receive a series of object trajectory and the corresponding timestamps of the coordinates in the trajectory. You can treat the trajectory as linestring.
Sensor A: [(x_1, y_1), (x_2, y_2), ..., (x_n, y_n)]
... | [0.0] |
0 |
Help me answer question regarding spatial relationship in a 2D plane:
Given Information:
You will receive a series of object trajectory and the corresponding timestamps of the coordinates in the trajectory. You can treat the trajectory as linestring.
Sensor A: [(x_1, y_1), (x_2, y_2), ..., (x_n, y_n)]
... | [1.0] |
0 |
Help me answer question regarding spatial relationship in a 2D plane:
Given Information:
You will receive a series of object trajectory and the corresponding timestamps of the coordinates in the trajectory. You can treat the trajectory as linestring.
Sensor A: [(x_1, y_1), (x_2, y_2), ..., (x_n, y_n)]
... | [1.0] |
0 |
Help me answer question regarding spatial relationship in a 2D plane:
Given Information:
You will receive a series of object trajectory and the corresponding timestamps of the coordinates in the trajectory. You can treat the trajectory as linestring.
Sensor A: [(x_1, y_1), (x_2, y_2), ..., (x_n, y_n)]
... | [1.0] |
0 |
Help me answer question regarding spatial relationship in a 2D plane:
Given Information:
You will receive a series of object trajectory and the corresponding timestamps of the coordinates in the trajectory. You can treat the trajectory as linestring.
Sensor A: [(x_1, y_1), (x_2, y_2), ..., (x_n, y_n)]
... | [0.0] |
0 |
Help me answer question regarding spatial relationship in a 2D plane:
Given Information:
You will receive a series of object trajectory and the corresponding timestamps of the coordinates in the trajectory. You can treat the trajectory as linestring.
Sensor A: [(x_1, y_1), (x_2, y_2), ..., (x_n, y_n)]
... | [0.0] |
0 |
Help me answer question regarding spatial relationship in a 2D plane:
Given Information:
You will receive a series of object trajectory and the corresponding timestamps of the coordinates in the trajectory. You can treat the trajectory as linestring.
Sensor A: [(x_1, y_1), (x_2, y_2), ..., (x_n, y_n)]
... | [0.0] |
0 |
Help me answer question regarding spatial relationship in a 2D plane:
Given Information:
You will receive a series of object trajectory and the corresponding timestamps of the coordinates in the trajectory. You can treat the trajectory as linestring.
Sensor A: [(x_1, y_1), (x_2, y_2), ..., (x_n, y_n)]
... | [1.0] |
End of preview. Expand in Data Studio
STARK: Spatial-Temporal reAsoning benchmaRK
STARK is a comprehensive benchmark designed to systematically evaluate large language models (LLMs) and large reasoning models (LRMs) on spatial-temporal reasoning tasks, particularly for applications in cyber-physical systems (CPS) such as robotics, autonomous vehicles, and smart city infrastructure.
Dataset Summary
Hierarchical Benchmark: Tasks are structured across three levels of reasoning complexity:
- State Estimation: Field variable prediction, spatial/temporal localization, and tracking with diverse sensor modalities (range, bearing, proximity, event-based).
- Reasoning Over Estimated States: Inference of spatial, temporal, and spatiotemporal relationships using formal logic frameworks (DE-9IM for space, Allen’s interval algebra for time).
- World-Knowledge-Aware Reasoning: Context- and knowledge-rich challenges such as intent prediction, route planning, landmark reasoning, and human mobility forecasting.
Scale and Diversity: Contains 25 unique tasks, over 10k challenge instances, and supports open-ended answers.
Sensor Modalities: Simulates real-world data from range sensors, bearing sensors, region (proximity) sensors, and event-based (e.g., TOA) sensors, as well as real and synthetic field variable datasets (e.g., air quality, traffic, temperature).
Evaluation Focus: Tasks are designed to assess both direct reasoning and the ability to generate and execute Python code. Baseline methods include classical geometric algorithms (multilateration, triangulation, Kalman filtering) and FSMs for human activity modeling.
Reproducibility: All data, code, and evaluation scripts are open-sourced to encourage benchmarking and method development in spatial-temporal reasoning. Please refer to our GitHub repo.
Example Use Cases
- Benchmarking LLM and LRM performance in geometric localization, trajectory tracking, spatial/temporal relationship inference, and real-world navigation tasks.
- Evaluating multi-step reasoning pipelines in simulated CPS environments.
- Assessing both direct question-answering and tool-use (Python code generation) capabilities.
How to use
- Untar the data_final.tar.gz file:
tar -xzvf data_final.tar.gz
- Rename the directory:
mv data_final_v5/ data
- Begin to use
python main.py --openai $m --dataset $t --index $i --mode $mode
Contact Information
If you have any questions or feedback, feel free to reach out:
- Name: Pengrui Quan
- Email: prquan@ucla.edu
Preprint
For more details, refer to our preprint.
License
Copyright (c) 2025, UCLA Networked and Embedded Systems Laboratory (NESL) All rights reserved.
Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:
Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.
Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution.
Neither the name of the copyright holder nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
- Downloads last month
- 2,544