toiar/khasi-instruction-response-v1

Dataset Card for Khasi Instruction-Response Dataset v1

Dataset Summary

Khasi Instruction-Response Dataset v1 is a curated collection of prompts and responses in the Khasi language, designed to support the fine-tuning of instruction-following language models. It includes tasks such as Q&A, translation, summarization, and culturally-grounded dialogue. The dataset reflects indigenous knowledge systems, cultural expressions, and educational content unique to the Khasi context, helping advance NLP in low-resource language settings.

Dataset Details

Dataset Sources

• Repository: https://huggingface.co/datasets/toiar/Khasi-Instruct-v1

Uses

Direct Use

This dataset is intended for:

• Fine-tuning instruction-tuned LLMs such as LLaMA, Mistral, Gemma
• Building Khasi-language conversational assistants or chatbots
• Research in indigenous knowledge systems and low-resource NLP

Out-of-Scope Use

• Commercial applications without explicit permission
• Misuse for harmful or culturally disrespectful generation tasks

Dataset Structure

Each data sample is a JSON object:

{
  "instruction": "Kaei ka Artificial Intelligence ?",
  "response": "Ka Artificial Intelligence (AI) ka dei ka jingstad ba la thaw da ki mashin ban leh ia ki kam kiba ju donkam ia ka jingstad u briew. Ka kynthup ia ki kam kum ka jinghikai (learning), ka jingpyrkhat (reasoning), bad ka jingpynbeit ia ki jingeh (problem-solving)."
}