A conversational Hinglish dataset built for fine-tuning Llama 3.1 8B.Hinglish = Hindi + English mixed in Roman script — exactly how people type on WhatsApp, Instagram, and daily chat.

📌 Dataset Description

This dataset contains single-turn instruction-response pairs in Hinglish (Roman script only).No Devanagari script — pure WhatsApp-style typing. The goal is to make Llama 3.1 8B respond naturally in Hinglish, the way a… See the full description on the dataset page: https://huggingface.co/datasets/nikhilnebhavani/hinglish-llama3-dataset.