Datasets:
Add 1 files
Browse files
README.md
ADDED
|
@@ -0,0 +1,62 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
annotations_creators: []
|
| 3 |
+
language:
|
| 4 |
+
- te
|
| 5 |
+
language_creators: []
|
| 6 |
+
license: []
|
| 7 |
+
multilinguality:
|
| 8 |
+
- monolingual
|
| 9 |
+
pretty_name: 'adaption-telugu_safety_prompts'
|
| 10 |
+
size_categories:
|
| 11 |
+
- 10K<n<100K
|
| 12 |
+
source_datasets:
|
| 13 |
+
- 'extended|https://huggingface.co/datasets/safety-aya/Aegis-AI-Content-Safety-Dataset-2.0-Telugu-safety'
|
| 14 |
+
tags:
|
| 15 |
+
- adaption
|
| 16 |
+
- instruction-tuning
|
| 17 |
+
- language
|
| 18 |
+
- writing-editing-communication
|
| 19 |
+
- personal-growth
|
| 20 |
+
task_categories: []
|
| 21 |
+
task_ids: []
|
| 22 |
+
---
|
| 23 |
+
|
| 24 |
+

|
| 25 |
+
|
| 26 |
+
This dataset is a remastered version of this [dataset](https://huggingface.co/datasets/safety-aya/Aegis-AI-Content-Safety-Dataset-2.0-Telugu-safety) prepared using [Adaption's](https://adaptionlabs.ai/app/auth) Adaptive Data platform.
|
| 27 |
+
|
| 28 |
+
# adaption-telugu_safety_prompts
|
| 29 |
+
|
| 30 |
+
A dataset of Telugu-language prompts labeled as SAFE or UNSAFE based on content safety. It includes text prompts and their corresponding safety annotations, intended for training or evaluating content moderation systems. The dataset focuses on identifying harmful or inappropriate content in Telugu.
|
| 31 |
+
|
| 32 |
+
### Dataset size
|
| 33 |
+
|
| 34 |
+
There are 13,920 data points in this dataset. This is an instruction tuning dataset.
|
| 35 |
+
|
| 36 |
+
### Quality of Remastered Dataset
|
| 37 |
+
|
| 38 |
+
The final quality is C, with a relative quality improvement of 116.7%.
|
| 39 |
+
|
| 40 |
+
### Domain
|
| 41 |
+
- Language (18%)
|
| 42 |
+
- Writing-editing-communication (12%)
|
| 43 |
+
- Personal-growth (10%)
|
| 44 |
+
|
| 45 |
+
### Language
|
| 46 |
+
- Telugu (100%)
|
| 47 |
+
|
| 48 |
+
### Tone
|
| 49 |
+
- Cautious (18%)
|
| 50 |
+
- Helpful (14%)
|
| 51 |
+
- Practical (10%)
|
| 52 |
+
|
| 53 |
+
### Evaluation Results
|
| 54 |
+
|
| 55 |
+
- **Quality Gains:**
|
| 56 |
+
<img src="https://proteus-prod-public.s3.us-east-1.amazonaws.com/temp/6aef9dfa-dbc2-4642-b027-01b4df9f7363.png" alt="QualityGains" style="max-width: 50%; display: block; margin-left: auto; margin-right: auto;" />
|
| 57 |
+
|
| 58 |
+
- **Grade Improvement:**
|
| 59 |
+
<img src="https://proteus-prod-public.s3.us-east-1.amazonaws.com/temp/f7329760-b0a7-4dd3-89e3-f1da39f24d6e.png" alt="Grade" style="max-width: 50%; display: block; margin-left: auto; margin-right: auto;" />
|
| 60 |
+
|
| 61 |
+
- **Percentile Chart:**
|
| 62 |
+
<img src="https://proteus-prod-public.s3.us-east-1.amazonaws.com/temp/820b967e-d6fc-41e3-a280-8248234c70e9.png" alt="Percentile Chart" style="max-width: 50%; display: block; margin-left: auto; margin-right: auto;" />
|