Étagère mesuréeIA et apprentissage automatique
Transformer explainers, ranked by where the analogy ends
- Remplissage médian
- 36%
- La partie utile commence
- 1:36
- Durée typique
- 29 min
dernier recalcul 16 sept. 2026
Every video on this shelf has to solve the same problem: attention is a mechanism made of matrix multiplication, and matrix multiplication is not something you can watch. So each of them picks a metaphor — a lookup table, a search engine, a room where words vote on each other — and the entire quality of the video is decided by what happens after the metaphor.
The ones that rank highly here spend the metaphor quickly and then show the actual operation: what the three projections are, what shape the things being multiplied have, why the scaling term is there, what the mask does. The ones lower down keep the metaphor going for forty minutes, restated in progressively more elaborate ways, and never show a number.
Both types are titled “transformers explained”. The density measurement is what separates them, because a metaphor restated is by definition not new information, and it is the single most common way this topic becomes an hour of video.
Entries carry a difficulty level, and on this topic it matters more than anywhere else on the site: a video that is perfect for someone who already knows what a dot product does is a wasted hour for someone who does not, and the reverse is worse.
-
01
Let's reproduce GPT-2 (124M)
0:004:01:26- Densité
- 84
- Remplissage
- 27%
- Utile à partir de
- 3:34
A code-complete, first-principles GPT-2 reproduction — one of the most valuable hands-on deep learning tutorials available.
-
02
Attention in transformers, step-by-step | Deep Learning Chapter 6
0:0026:10- Densité
- 83
- Remplissage
- 21%
- Utile à partir de
- 1:40
A masterfully clear, dense walkthrough of the attention mechanism that rewards careful watching with real technical understanding.
-
03
Transformers, the tech behind LLMs | Deep Learning Chapter 5
0:0027:14- Densité
- 80
- Remplissage
- 26%
- Utile à partir de
- 1:26
A masterclass primer on transformer internals — dense, rigorous, and exactly what its title promises.
-
04
Let's build GPT: from scratch, in code, spelled out.
0:001:56:20- Densité
- 82
- Remplissage
- 35%
- Utile à partir de
- 14:11
A masterclass build-along: real working GPT code, the actual mechanics behind ChatGPT, with almost no filler.
-
05
Transformer Neural Networks, ChatGPT's foundation, Clearly Explained!!!
0:0036:15- Densité
- 79
- Remplissage
- 23%
- Utile à partir de
- 1:20
A rigorous, worked-numbers walkthrough of transformer internals that actually teaches how ChatGPT-style models work, not just what they do.
-
06
Large Language Models explained briefly
0:007:58- Densité
- 75
- Remplissage
- 26%
- Utile à partir de
- 0:33
A masterfully compressed, accurate primer on how LLMs actually work — dense, honest, no sales pitch.
-
07
Transformers Step-by-Step Explained (Attention Is All You Need)
0:0010:04- Densité
- 70
- Remplissage
- 39%
- Utile à partir de
- 2:40
A tight, genuinely educational explainer of Transformer attention with a real worked example, lightly interrupted by a disclosed sponsor read.
-
08
Transformer Neural Networks - EXPLAINED! (Attention is all you need)
0:0013:05- Densité
- 68
- Remplissage
- 31%
- Utile à partir de
- 1:54
A dense, well-structured conceptual walkthrough of transformer architecture that earns its 'EXPLAINED' title with zero filler.
-
09
Illustrated Guide to Transformers Neural Network: A step by step explanation
0:0015:01- Densité
- 72
- Remplissage
- 31%
- Utile à partir de
- 1:03
A tight, accurate conceptual tour of the Transformer's internals — solid teaching, though it retreads familiar illustrated-guide territory rather than breaking new ground.
-
10
Transformers explained | The architecture behind LLMs
0:0019:48- Densité
- 71
- Remplissage
- 33%
- Utile à partir de
- 0:38
A genuinely dense, accurate transformer explainer that earns its title with real mechanics, not hype.
-
11
Transformers: The best idea in AI | Andrej Karpathy and Lex Fridman
0:008:38- Densité
- 69
- Remplissage
- 34%
- Utile à partir de
- 2:44
A sharp, dense breakdown of why the Transformer works — one of the clearer plain-language explanations of its design philosophy, if brief.
-
12
Transformers, explained: Understand the model behind GPT, BERT, and T5
0:009:11- Densité
- 67
- Remplissage
- 34%
- Utile à partir de
- 1:25
A genuinely solid, jargon-light explainer of transformer architecture that earns its title without ever really selling anything.
-
13
Transformers for beginners | What are they and how do they work
0:0019:59- Densité
- 68
- Remplissage
- 29%
- Utile à partir de
- 0:31
Solid, math-grounded beginner explainer of transformer internals, lightly bookended by the channel's own API plug.
-
14
Transformers Explained | Simple Explanation of Transformers
0:0057:31- Densité
- 67
- Remplissage
- 27%
- Utile à partir de
- 1:36
A patient, analogy-heavy but genuinely thorough walkthrough of Transformer internals — worth the long runtime if you already know your deep learning basics.
-
15
How does AI actually work? Transformers explained
0:0032:21- Densité
- 66
- Remplissage
- 30%
- Utile à partir de
- 0:31
A solid, honestly-titled conceptual explainer of Transformer architecture, weakened only by redundant recaps and a mid-video sponsor detour.
-
16
Transformer Architecture Explained 'Attention Is All You Need'
0:0012:49- Densité
- 62
- Remplissage
- 34%
- Utile à partir de
- 0:47
A clear, well-paced conceptual primer on Transformer attention — not groundbreaking, but a genuinely solid explainer worth the 13 minutes for newcomers to the architecture.
-
17
Transformers, explained: Understand the model behind ChatGPT
0:0024:07- Densité
- 61
- Remplissage
- 28%
- Utile à partir de
- 3:29
A clear, accessible mental model of how Transformers work end-to-end — skips the real attention math (no Q/K/V) but is genuinely useful as a conceptual primer.
-
18
What are Large Language Models (LLMs)?
0:005:30- Densité
- 56
- Remplissage
- 41%
- Utile à partir de
- 0:32
A tight, honest beginner explainer of LLMs and prompt design — light on depth but dense and accurate for its length.
-
19
Everything You Need To Know About Large Language Models (LLMs)
0:0025:20- Densité
- 58
- Remplissage
- 39%
- Utile à partir de
- 0:32
A solid, broad beginner's overview of LLM mechanics and history, only lightly diluted by a sponsor segment for AI Camp.
-
20
The Transformer architecture
0:002:45- Densité
- 50
- Remplissage
- 47%
- Utile à partir de
- 0:59
A clean, honest, high-level primer that sets up the series without pretending to teach the deep mechanics yet.
-
21
What are Transformers (Machine Learning Model)?
0:005:51- Densité
- 51
- Remplissage
- 46%
- Utile à partir de
- 1:17
A clear, accurate but fairly standard conceptual primer on transformers — solid intro, low novelty.
-
22
Transformer Explained
0:006:55- Densité
- 55
- Remplissage
- 40%
- Utile à partir de
- 2:07
A solid conceptual primer on transformer limitations and fixes, but it openly admits it skips the actual mechanics the title implies.
-
23
Large Language Models Explained Simply (In 13 Minutes)
0:0012:57- Densité
- 50
- Remplissage
- 45%
- Utile à partir de
- 3:21
A clear, if conceptually shallow, LLM 101 explainer padded with light jokes and capped by a short affiliate plug.
-
24
Large Language Models | How Large Language Models Work? | Introduction to LLM | Simplilearn
0:0015:47- Densité
- 46
- Remplissage
- 48%
- Utile à partir de
- 2:46
A solid, if generic, beginner overview of how LLMs and transformers work, padded with a short in-house course pitch.
-
25
How Large Language Models Work
0:005:34- Densité
- 48
- Remplissage
- 51%
- Utile à partir de
- 2:04
A clear, competent beginner overview of LLM mechanics from IBM, though fairly generic and light on real depth.
Gistil's own measurement. Not a YouTube rating, and not the channel's position. Le classement est le nôtre ; les vidéos appartiennent à leurs chaînes.
Comment ce classement est établi
Chaque vidéo ici a été mesurée sur les mêmes axes — quelle part de sa durée porte de l'information, quelle part est du remplissage, et à quelle seconde elle commence à payer. L'ordre suit la densité de valeur, pas les vues, la fraîcheur ou à quel point on a aimé la chaîne. Ce que signifient les chiffres →
← Toutes les vidéos mesurées en ia et apprentissage automatique