metadata
annotations_creators: []
language:
- te
language_creators: []
license: []
multilinguality:
- monolingual
pretty_name: adaption-telugu_normalization
size_categories:
- 1K<n<10K
source_datasets:
- original
tags:
- adaption
- instruction-tuning
- legal
- technology
- governance
task_categories: []
task_ids: []
This dataset is a remastered version prepared using Adaption's Adaptive Data platform.
adaption-telugu_normalization
This dataset contains a collection of user queries and statements written in Telugu, covering diverse topics such as mobile troubleshooting, cyber security threats, insurance renewals, and emergency services. The samples vary from short keywords to detailed problem descriptions involving scams, natural disasters, and administrative procedures. It represents real-world information seeking behavior and support requests from Telugu-speaking users.
Dataset size
There are 8,174 data points in this dataset. This is an instruction tuning dataset.
Quality of Remastered Dataset
The final quality is B, with a relative quality improvement of 40.0%.
Domain
- Legal (18%)
- Technology (16%)
- Governance (14%)
Language
- Telugu (100%)
Tone
- Helpful (36%)
- Informative (32%)
- Urgent (16%)
Evaluation Results
Quality Gains:
Grade Improvement:
Percentile Chart:
