C-Eval: vivo introduces large-scale cloud solutions, with indicators approaching GPT-4

According to news from this site on August 15, vivo plans to strengthen its strength in the field of large models. After Huawei Xiaoyi, Xiaomi Xiaoai classmates, and OPPO Xiaobu Assistant upgrade their AI large model capabilities

Huawei, Xiaomi , OPPO has started or is about to start testing its related large models and new intelligent assistants, and vivo's self-developed large model cloud solution has also appeared on the C-Eval rankings. It is expected that in OriginOS 4, Jovi Assistant will usher in some unique features

Note from this site: C-Eval is a multi-item command containing 13948 A comprehensive Chinese basic model assessment suite of multiple-choice questions, covering four major directions: humanities, social sciences, science and engineering, and other majors, including a total of 52 subjects and four difficulty levels

According to the information provided by C-Eval, vivo_Agent_LM_7B is a large-scale pre-trained language model independently developed by vivo AI Global Research Institute. Its name indicates that it has 7 billion parameters

According to the final results, vivo_Agent_LM_7B scored 64.4 points, ranking fifth. It is almost the same as the second-place GPT-4, and even performs better in the humanities field, and also performs well in the social science field

Read Related information:

Lei Jun announced that Xiaomi will form a large team in April, and the new version of Xiao Ai will be invited to test from today

OPPO Xiaobu Assistant will be based on AndesGPT and has a powerful AI system Model capabilities

Huawei Yu Chengdong said that the large AI model will provide support for the new Xiaoyi HarmonyOS 4, emphasizing the insistence on AI assistance rather than replacement

The article on this site contains external jump links for Provide more information and save screening time. The results are for reference only

The above is the detailed content of C-Eval: vivo introduces large-scale cloud solutions, with indicators approaching GPT-4. For more information, please follow other related articles on the PHP Chinese website!