Professor Wang Benyou's team from The Chinese University of Hong Kong, Shenzhen (CUHK-Shenzhen) and Shenzhen Institutes of Big Data (SIBD) has developed open-source large language models called Phoenix and Chimera.
The Chinese language capabilities of these models are comparable to Baidu Wenxin Yiyan, and in the evaluation of GPT-4, they achieved a 97% level of performance compared to Wenxin Yiyan. In human evaluations, they are on par with Wenxin Yiyan in 50% of cases.
The GitHub repository for maintaining Phoenix and Chimera models is named "LLM Zoo," which also includes other upcoming models such as Huatuo GPT for the medical field (huatuogpt.cn), legal GPT models, multimodal large models, and retrieval-enhanced large models. For more details, please refer to the technical report on their GitHub homepage.