嵌入层Embedding Layer
Embedding Layer
解决什么问题What problem it solves
把词变成数字向量的第一步，模型内部只认数字。The first step of turning words into numeric vectors — models only work in numbers internally.
这是什么What it is
模型内部把每个 token 映射成高维向量的那一层，是向量检索技术的源头。
The layer mapping each token to a high-dimensional vector inside the model — the origin of vector retrieval.
为什么需要知道Why it matters
它连接「看懂站」和「用深站」。你理解的 embedding 概念，和这里的实现是同一个东西。It connects the conceptual layer to the applied one: the embedding you use in practice is this.
现在能做什么What you can do now
回到名词地图看第4站的其他 35 个词 →Back to the map: the other 35 terms in station 4 →
出处Sources
- Word2Vec（论文）Word2Vec: Efficient Estimation of Word Representations ↗