Deep Dive into LLMs like ChatGPT
10 人推荐 · 跨 8 个月 · 2025–2026
中文导读
Andrej Karpathy 长达 3 个多小时的通俗硬核大课，全面拆解类 ChatGPT 大模型的构建全貌：从预训练到后训练全链路，并帮大家建立理解幻觉等特性的心智模型，掌握用好 LLM 的诀窍。
为什么 HN 推它
HN 上推荐频率最高的大模型入门讲座，几乎不堆晦涩数学，却把 LLM 原理讲得通透接地气。
「强烈推荐 Andrej Karpathy 这期讲 ChatGPT 类大模型的深度解析。」HN 网友原话：I highly recommend Andrej Karpathy's Deep Dive into LLMs like ChatGPT
英文原摘要（HackerTube / Claude 整理）
Andrej Karpathy's three-hour-plus general-audience deep dive into how ChatGPT-style LLMs are built, covering the full training stack from pretraining to post-training. It also gives mental models for their quirks, like hallucination, and how to use them well.
HN's most frequent recommendation for a clear, approachable foundation in how LLMs work without heavy math.
HN 上的提及
he mentioned how hard it is for LLMs to have humor and be funny
He goes over a ton of different stuff related to model training, but makes it very easy to understand.
He said that hallucination was because in the training data there were no examples where the answer is, "I don't know".
Karpathy has a good intro video on this.
There is a very good intro on how LLM work by Kharpaty.