---
license: apache-2.0
dataset_info:
  features:
  - name: question_id
    dtype: string
  - name: category
    dtype: string
  - name: cluster
    dtype: string
  - name: turns
    list:
    - name: content
      dtype: string
  splits:
  - name: train
    num_bytes: 251691
    num_examples: 500
  download_size: 154022
  dataset_size: 251691
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---

## Arena-Hard-Auto

**Arena-Hard-Auto-v0.1** ([See Paper](https://arxiv.org/abs/2406.11939)) is an automatic evaluation tool for instruction-tuned LLMs. It contains 500 challenging user queries sourced from Chatbot Arena. We prompt GPT-4-Turbo as judge to compare the models' responses against a baseline model (default: GPT-4-0314). Notably, Arena-Hard-Auto has the highest *correlation* and *separability* to Chatbot Arena among popular open-ended LLM benchmarks ([See Paper](https://arxiv.org/abs/2406.11939)). If you are curious to see how well your model might perform on Chatbot Arena, we recommend trying Arena-Hard-Auto.

Please checkout our GitHub repo on how to evaluate models using Arena-Hard-Auto and more information about the benchmark. 

If you find this dataset useful, feel free to cite us!
```
@article{li2024crowdsourced,
  title={From Crowdsourced Data to High-Quality Benchmarks: Arena-Hard and BenchBuilder Pipeline},
  author={Li, Tianle and Chiang, Wei-Lin and Frick, Evan and Dunlap, Lisa and Wu, Tianhao and Zhu, Banghua and Gonzalez, Joseph E and Stoica, Ion},
  journal={arXiv preprint arXiv:2406.11939},
  year={2024}
}
```