[1hr Talk] Intro to Large Language Models
9 人推荐 · 跨 6 个月 · 2023–2026
![[1hr Talk] Intro to Large Language Models](https://i.ytimg.com/vi/zjkBMFhNj_g/hq2.jpg)
中文导读
Andrej Karpathy 用一小时为大众系统梳理了大语言模型（LLM）：从基本工作原理、预训练与微调，到将 LLM 类比为操作系统的独到视角，再到越狱等安全挑战，内容极其硬核凝练。
为什么 HN 推它
HN 社区公认最扎实通俗的 LLM 入门指南，社区在讨论模型缩放（scaling）与训练细节时经常直接引用该视频。
「先看他这个 LLM 快速入门，虽然只有一小时，但全是高密度的干货。」HN 网友原话：Start with his short intro to LLMs, only 1 hour long but is dense with valuable stuff
英文原摘要（HackerTube / Claude 整理）
Andrej Karpathy gives a one-hour general-audience introduction to large language models, covering what they are, how they are trained, comparisons to operating systems, and security challenges such as jailbreaks.
HN recommends it as the dense, approachable starting point for understanding LLMs and often cites it for specific claims about scaling and training.
HN 上的提及
Andrej Karpathy mentions it in his new talk (~38 minutes in)
Serendipitously, Karpathy broadly discussed these two issues yesterday !
Karpathy in his recent video [1] agrees
it's a very nice bonus, but we can sort of get more powerful models for free
an amazing high-level overview of LLMs, including extensive discussion about attack vectors