NVIDIA Nemotron: Korean Personas (1M Synthetic Personas)

1. Open-Source Basic Metadata Asset Name: NVIDIA Nemotron-Personas-Korea (1M Synthetic Personas) Category: Synthetic Users / AI Prompts / Market Research Open-Source Provider: NVIDIA (Curated by SampleLake) Link: Access on Hugging Face (https://huggingface.co/datasets/nvidia/Nemotron-Personas-Korea) 2. Overview "1 Million High-Fidelity Synthetic Personas Mirrored After Real-World South Korean Demographics." Nemotron-Personas-Korea is a 1-million-row synthetic dataset open-sourced by NVIDIA. Unlike generic mock data, it strictly aligns with official South Korean demographic distributions from Statistics Korea (KOSIS) and the Supreme Court—including precise ratios for regions, age, gender, and occupation. 3. Key Features & Use Cases Multi-Dimensional Depth: Spans 1.7 billion tokens across 7 distinct aspects per profile (Career, Family, Lifestyle, Culinary Tastes, etc.).