OpenBMB releases bilingual UltraData-RL dataset for RL training

openbmb · hf · 2026-09-08

OpenBMB's UltraData-RL-2609 dataset is trending on Hugging Face: a bilingual (English/Chinese) text-generation and QA dataset of 10K-100K rows under apache-2.0, tied to two arXiv papers and aimed at RL training data.

Original post →

More from Research

Research channel →