Feed Entities Podcasts Deep Dives Developers About

[KongchangAI]

Real-time AI knowledge base. Every claim is traceable to its source, with data freshness under 24 hours.

Powered by 240+ first-hand AI sources

Knowledge

Feed
Hot Entities
Podcasts

Developers

CLI
MCP Server
REST API
About

© 2026 KongchangAI. The freshest, most accurate AI knowledge

Unverified50% confidenceFactExact time

语言模型生成 token 本质上是内存带宽受限任务而非纯算力受限

1

Sources

50%

Confidence

Long-term

Relevance

7/8/2026

First Seen

Sources

GPT-5.6 Sol限量预览：分级模型Sol/Tera/Luna与准入机制详解

bilibili开水庆余年6/27/2026

Related Claims

Unverified大语言模型推理中输出 token 的实际硬件成本远高于输入 token，因解码阶段是内存带宽受限的串行过程81% similar Verified大语言模型的推理过程本质上是内存带宽受限（memory-bandwidth bound）的任务，每生成一个Token都需要将模型的全部权重从存储中读取一遍79% similar VerifiedToken是大语言模型处理文本的基本计量单位，由BPE（字节对编码）等分词算法生成75% similar Unverifiedtoken是大模型计费的依据，由分词器将原始文本切分而来，是大模型处理文本的基本单位74% similar Unverified当前主流大语言模型几乎都采用自回归方式逐 token 顺序生成文本74% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/275716

API

curl https://kongchang.com/api/v1/knowledge/claims/275716

MCP

get_claim(id=275716)