---
license: apache-2.0
tags:
- planning
- natural-language
- trip-planning
- meeting-scheduling
- llm-evaluation
language:
- en
size_categories:
- 1K<n<10K
---

# naturalplan-benchmark

Natural Plan: Benchmarking LLMs on Natural Language Planning (Trip, Meeting, Calendar)

## Dataset Description

This dataset is a standardized version of the original benchmark, prepared for easy evaluation of LLMs on planning tasks.

### Splits

- **test**: 3,600 samples


### Features

```python
{
  "id": "Value(dtype='string', id=None)",
  "task_type": "Value(dtype='string', id=None)",
  "prompt_0shot": "Value(dtype='string', id=None)",
  "prompt_5shot": "Value(dtype='string', id=None)",
  "golden_plan": "Value(dtype='string', id=None)",
  "num_cities": "Value(dtype='string', id=None)",
  "cities": "Value(dtype='string', id=None)",
  "durations": "Value(dtype='string', id=None)"
}
```

## Usage

```python
from datasets import load_dataset

# Load the dataset
dataset = load_dataset("tuandunghcmut/naturalplan-benchmark")

# Access splits
train_data = dataset["train"]  # if available
test_data = dataset["test"]

# Example
print(test_data[0])
```

## Citation

If you use this dataset, please cite the original paper:

```bibtex
@article{zheng2024naturalplanbenchmarkingllms,
  title={NATURAL PLAN: Benchmarking LLMs on Natural Language Planning},
  author={Huaixiu Steven Zheng and Swaroop Mishra and Hugh Zhang and Xinyun Chen and Minmin Chen and Azade Nova and Le Hou and Heng-Tze Cheng and Quoc V. Le and Ed H. Chi and Denny Zhou},
  year={2024},
  eprint={2406.04520},
  archivePrefix={arXiv
}
```

## Acknowledgments

This dataset is a reformatted version of the original benchmark. 
All credit goes to the original authors.

---

**Prepared by**: tuandunghcmut
**Original Source**: See citation above