You have visual language understanding that you just showed.

可以具备刚才展示的视觉语言理解能力。

And when you're learning a language, understanding the nuances of context and tone and body language are so, so important.

当你学习一门语言的时候，理解语境、语调以及身体语言的细微差别也非常非常重要。

So, here's a simple example of the unconscious in modern psychology: Language understanding.

所以，这是现代心理学中无意识的一个简单例子：语言理解。

To measure understanding, we are sampling from the massive multitask language understanding benchmark.

为了衡量理解程度，我们从大规模多任务语言理解基准中进行抽样。

One problem with benchmarks such as MMLU (massive multi-task language understanding) is that they are simply too easy for today's models.

MMLU（大规模多任务语言理解）等基准测试的一个问题是，它们对于当今的模型来说太简单了。

Siri also has richer language understanding, meaning you can speak to it more naturally and it will know what you're talking about.

Siri 还具有更丰富的语言理解能力，这意味着你可以更自然地与它交谈， 它也会知道你在说什么。

And when we're speaking in another language, understanding and tuning into that song or that intonation can be just as important as understanding the pronunciation and the word stress.

当我们用另一种语言说话时，理解和调整音调或语调与理解发音和单词重音同样重要。

The company has explored deep into various fields, including AI models, machine learning platforms, machine learning cloud platforms, AIGC, natural language understanding, and digital humans.

在AI模型方面，腾讯开发了混元AI大模型，覆盖NLP、CV和多模态，参数规模从亿级别到万亿级别。该公司在多项基准测试中表现优异，包括CLUE和VCR，已经成功落地于腾讯广告、搜索、对话等内部产品和腾讯云服务的外部客户。腾讯还在不断探索更大的模型和提高在文本内容生成、图像识别等方面的能力。