XuanYuan2 70B

About

The XuanYuan2-70B is an advanced large language model built on the XuanYuan-70B base, featuring enhanced capabilities through vast high-quality data pre-training and instruction fine-tuning. It employs reinforcement learning with human feedback (RLHF) to better align with human preferences, offering improved performance, safety, and notable proficiency in financial tasks. The model supports an extended context length of 16k tokens, improving its ability to handle long-form text. It belongs to a family of models that includes a chat variant and quantized versions, leveraging Llama2 architecture to excel in natural language processing, particularly for financial applications 5.

Capabilities