Skip to content
Trending storyObserving

LLM Token 机制、上下文窗口与 API 计费

1 report1 reporting sourceUpdated 1 day ago

Understand the story

AI synthesis

围绕大语言模型如何处理输入与计费,Tproger 的报道解释了 token 的基本机制:文本在进入语言模型前会被 tokenizer 切分成 token,并映射为词表编号。报道强调 token 并不等于单词,也不是固定字符数,例如 BERT 的 tokenizer 会把 "I have a new GPU!" 切成多个 token。该报道同时涉及 token 如何影响大模型请求的成本,以及上下文窗口限制与 API 计费方式。目前进展停留在机制层面的科普说明,尚未出现新的模型、工具或计费政策变动。

Generated by AI from reports · Updated 9 hours ago

Report timeline

Follow the reports to explore different perspectives.

10/5
  1. Tproger · Программирование
    什么是 token,以及它如何影响大模型请求的成本

    文本进入语言模型前会被 tokenizer 切成 token 并映射为词表编号,token 不等于单词或固定字符数,例如 BERT 的 tokenizer 会把 "I have a new GPU!

Interest in this story

Current interest 4·Peak within comparable coverage 6(Oct 6, 16:00)·Change over 24 hours within comparable coverage –

024610/616:0010/619:0010/621:0010/700:00

The trend compares the same participants under continuous, complete observation, so its coverage may be narrower than the current score. Hover or tap to view hourly interest; use the left and right arrow keys to navigate.