China's Baichuan Intelligent Technology Unveils Open-Source 13B Parameter Large Language Model
www.maginative.com
The model has set a new benchmark, outstripping Meta's LLaMa-13B by a substantial 40% in training data volume.