商汤科技发布日日新SenseNova 4.0，大模型开启AI新体验
北京时间2023年3月8日，商汤科技正式发布“日日新SenseNova 4.0”，标志着其大模型体系迎来全面升级。
日日新SenseNova 4.0在多维度实现全面提升：
更全面的知识覆盖：SenseNova 4.0拥有海量的数据训练基础，覆盖了广泛的知识领域，包括自然语言、图像、视频、音频等。
更可靠的推理能力：通过引入先进的推理算法和知识图谱，SenseNova 4.0能够更准确、更可靠地进行推理和判断。
更优越的长文本理解力：SenseNova 4.0采用Transformer架构，拥有强大的长文本理解能力，能够有效处理复杂的长文本信息。
更稳定的数字推理能力：SenseNova 4.0在数字推理任务中表现出色，能够稳定地解决数学、逻辑等问题。
更强的代码生成能力：SenseNova 4.0具备强大的代码生成能力，能够根据自然语言描述自动生成代码，极大地提高了开发效率。
此外，日日新SenseNova 4.0还支持跨模态交互，能够无缝衔接不同模态的数据，实现更丰富的交互体验。
值得一提的是，日日新·商量大语言模型-通用版本（SenseChat V4）是SenseNova 4.0的重要组成部分。SenseChat V4支持128K语境窗口长度，综合整体评测成绩水平比肩GPT 4，相较GPT 3.5已经实现全面超越。
商汤科技表示，日日新SenseNova 4.0将广泛应用于智能客服、内容创作、教育培训、医疗健康等领域，为各行各业带来全新的AI体验。
英语如下：
**Headline:** SenseTime’s Large Model Gets Another Upgrade, Rivaling GPT-4
**Keywords:** Large model, AI experience, Language model
**News Content:** SenseTime Releases SenseNova 4.0, Large Model OpensNew AI Experience
Beijing Time, March 8, 2023, SenseTime Technology officially released “SenseNova 4.0”, marking a comprehensive upgrade of its large model system.
SenseNova 4.0 has achieved comprehensive improvements in multiple dimensions:
More comprehensive knowledge coverage: SenseNova4.0 has a massive data training foundation, covering a wide range of knowledge fields, including natural language, images, videos, audio, etc.
More reliable reasoning ability: By introducing advanced reasoning algorithms and knowledge graphs, SenseNova 4.0 can reason and judge more accurately and reliably.
Superior long text comprehension: SenseNova 4.0 adopts the Transformer architecture and has a strong long text comprehension ability, which can effectively process complex long text information.
More stable digital reasoning ability: SenseNova 4.0 performs well in digital reasoning tasks and can stably solve problems in mathematics, logic, etc.
Stronger code generation ability: SenseNova 4.0 has a strong code generation ability and can automatically generate code based on natural language descriptions, greatly improving development efficiency.
In addition, SenseNova 4.0 also supports cross-modal interaction, which can seamlessly connect data from different modalities to achieve a richer interactive experience.
It is worth mentioning that SenseTime·Shangliang Large Language Model-Universal Version (SenseChat V4) is an important part of SenseNova 4.0. SenseChat V4 supports a context window length of 128K, and its comprehensive evaluation results are comparable to GPT-4, surpassing GPT 3.5.
SenseTime Technology said that SenseNova 4.0 will be widely used in intelligent customer service, content creation, education and training, medical health, and other fields, bringing a new AI experience to all walks of life.
【来源】https://tidenews.com.cn/news.html?id=2705294&source=1
Views: 6