---
license: cc-by-nc-4.0
---
# RWQ Dataset

This dataset containes 20,772 questions to ensure that the evaluation of LLMs accurately reflects their use in practical scenarios, we assemble a dataset called "Real-World Questions" (RWQ). It comprises authentic questions sourced from various platforms such as [Google Trends](https://trends.google.com/trends/)(Transforming Google Trends entries into formulated questions with GPT-4), [Quora](https://www.quora.com/), [ShareGPT](https://sharegpt.com/), [LMSYS-Chat-1M](https://huggingface.co/datasets/lmsys/lmsys-chat-1m), and [AlpacaEval](https://huggingface.co/datasets/tatsu-lab/alpaca_eval).

## Basic Statistics

| source  |   | 
|---|---|
| Google Trend | 26.0% |
| LMSYS-Chat-1M | 24.6% |
| ShareGPT | 24.2% |
| Quora | 21.3% |
| AlpacaEval | 3.9% |

## Citation
TOOD