Recently, CTTL-Terminal Labs issued the "End-Side Large Language Model Capability Certificate" to the "vivo Blue Heart Edge-side Large Model V1.0" of vivo Mobile Communications Co., Ltd. According to FT-Z18-0113-01 "Evaluation Scheme for the Comprehensive Ability of Edge-side Large Language Models", 38 test cases in 4 dimensions of the edge-side large language model, including model basic ability, model benchmark performance, model service performance, and model application ability, are thoroughly tested and comprehensively evaluated. After evaluation, the "vivo blue heart edge-side large model V1.0" of vivo Mobile Communications Co., Ltd. passed the test of all service capability domains in the evaluation process, and this certificate is hereby issued.

Large language model is a natural language processing model based on deep learning, which aims to understand, generate and translate human language, and has shown great potential and value in improving human-computer interaction, information acquisition, content generation, and educational assistance. The large language model is trained with massive text data, and can learn rich semantic information and contextual relationships, which can be applied to text generation, question answering systems, translation and transformation, dialogue systems, etc., bringing users the experience of generating natural and fluent text, understanding the context and generating answers to specific contexts, the ability to handle multi-language tasks, and natural conversations. The edge-side large language model is becoming an important direction for the development of smart devices by improving response speed, protecting data privacy, and improving energy efficiency and user experience. With the continuous progress of edge computing and AI technology, its application prospects will become more and more extensive, which will have a profound impact on the intelligent society of the future.

In the future, the CTTL-Terminal Labs will continue to launch more evaluation programs, update and iterate the evaluation programs, ensure the effectiveness and reliability of the product in a professional and fair evaluation way, lay a solid foundation for its wide application, provide the public with rich consumer references and suggestions, and guide the healthy and sustainable development of the industry.

Consultation Contact:

Mr. Dong dongqianzhou@caict.ac.cn

Mrs. Huang huangyunxia@caict.ac.cn