File size: 2,922 Bytes
1b6139c db5748a 1b6139c db5748a |
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 |
---
dataset_info:
features:
- name: id
dtype: string
- name: prompt
dtype: string
- name: chosen
list:
- name: content
dtype: string
- name: role
dtype: string
- name: rejected
list:
- name: content
dtype: string
- name: role
dtype: string
- name: model
dtype: string
splits:
- name: train
num_bytes: 21292505
num_examples: 9500
download_size: 11187068
dataset_size: 21292505
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
license: apache-2.0
language:
- ar
tags:
- dpo
- orpo
---
# 🤗 Dataset Card for "NoRobots-Aya.23.8B-DPO"
### Dataset Sources & Infos
- **Data Origin**: Derived from the Arabic No Robots dataset : [2A2I/H4_no_robots](https://huggingface.co/datasets/2A2I/H4_no_robots) which is based on the original [No Robots](https://huggingface.co/datasets/HuggingFaceH4/no_robots) dataset inspired from the [InstructGPT](https://huggingface.co/papers/2203.02155) paper.
- **Languages**: Modern Standard Arabic (MSA)
- **License:** CC BY-NC 4.0
- **Maintainers:** [Ali Elfilali](https://huggingface.co/Ali-C137) and [Marwa El Kamil](https://huggingface.co/maghwa)
### Purpose
`NoRobots-Aya.23.8B-DPO` is a DPO dataset designed to advance Arabic NLP by comparing human-generated responses(translated), labeled as "chosen," with AI-generated responses, marked as "rejected." This approach helps improve the performance of Arabic language models by guiding them to produce more human-like and contextually appropriate responses.
### Usage
This dataset can be used to train and evaluate Arabic NLP models, particularly in tasks requiring nuanced language understanding and generation. By utilizing this dataset, researchers and developers can refine AI models to better distinguish between high-quality, human-like responses and less effective AI-generated ones, leading to more accurate and contextually relevant language models.
#### Use with HuggingFace
To load this dataset with Datasets, you'll need to install the datasets library with:
```bash
pip install datasets --upgrade
```
and then use the following code:
```python
from datasets import load_dataset
dataset = load_dataset("2A2I/NoRobots-Aya.23.8B-DPO")
```
### Contribution and Collaborative Engagement
Find 'NoRobots-Aya.23.8B-DPO' on the Hugging Face Hub at [2A2I/NoRobots-Aya.23.8B-DPO](https://huggingface.co/datasets/2A2I/NoRobots-Aya.23.8B-DPO), where community contributions are welcomed. Users are invited to share feedback and propose enhancements.
### Support and Collaborate
We are dedicated to cultivating an inclusive and encouraging space for Arabic AI and NLP research. For assistance, collaboration opportunities, or inquiries related to the dataset, please connect with us through the Hugging Face Hub's discussion section or contact us via [2A2I Contact Email](arabic.ai.initiative@gmail.com).😀
|