跳到主要内容
Learn Motion
探索课程产品能力价格

下载 Learn Motion

用手机扫描二维码,下载 App。

App Store

扫码从 App Store 下载 Learn Motion打开商店页面

Google Play

扫码从 Google Play 下载 Learn Motion打开商店页面
登录
Learn Motion

How ChatGPT Writes One Token at a Time

1From Your Words to Tokens2Predicting the Next Token3Choosing a Token and Looping4Why the Answer Reads as Coherent Text
Predicting the Next Token

One Score for Every Candidate

1 / 4
The prompt is now a row of token IDs sitting inside the context window. The model reads that row and does one thing: it asks which token should come next.
0:00 / 0:00

Given the token sequence so far, the model produces a raw score, called a logit, for every entry in its vocabulary. The vocabulary is fixed and large, so the output is a full row of scores, one per possible next token. These logits are not yet probabilities; they are unbounded numbers that can be positive or negative, and only their relative size matters at this stage.

上一页1 / 4下一页

Learn Motion

生成课程,认真学会。

运营方:Wuhan Daoyin Technology Co., Ltd.

联系我们: [email protected]
隐私政策用户服务协议

© 2026 Learn Motion