After Huawei Xiaoyi, Xiaomi Xiaoai, and OPPO Xiaobu Assistant upgraded their AI large-scale model capabilities, Vivo is also preparing to make efforts in the field of large-scale models.
At present, large models related to Huawei, Xiaomi, OPPO and new smart assistants have started or are about to start testing, and now, Vivo’s self-developed large model cloud solution has also appeared on the C-Eval ranking list, and may be included in the OriginOS 4 for Jovi Assistants make a difference.
Note: C-Eval is a comprehensive Chinese basic model evaluation kit. It contains 13,948 multiple-choice questions, covering four major directions of humanities, social sciences, science and engineering, and other majors, including 52 different subjects and four difficulty levels.
Judging from the information given by C-Eval, Vivo_Agent_LM_7B is a large-scale pre-trained language model independently developed by the Vivo AI Global Research Institute. It is not difficult to see from the naming that it has 7 billion parameters.
Just looking at the final result, the final score of Vivo_Agent_LM_7B is 64.4 points, ranking fifth in C-Eval, which is very close to the second-placed GPT-4, even stronger in the field of humanities, and also has a good performance in the field of social sciences.