NVIDIA and Elith Release Japanese Finance Persona Dataset

NVIDIA and Elith have published Nemotron-Personas-Finance-Japan on Hugging Face, a synthetic dataset of 9M personas (1M records with nine personas each) written in natural Japanese. It extends the open Nemotron-Personas-Japan dataset with household finance and financial literacy fields, built from official Japanese statistics and public survey data using NVIDIA's NeMo SDG-PGMs and NeMo Data Designer. Each record has 50 fields, and the release covers roughly 950k unique names and more than 1,600 occupation categories. It is available under CC BY 4.0 for commercial and non-commercial use. NVIDIA says the design incorporated feedback from Japanese financial institutions including Sompo Japan, Tokio Marine, Bank of Fukuoka, Mizuho, Meiji Yasuda and the FDUA. A companion dataset, Nemotron-PII-Japan, launched the same day.