Master AI Concepts
Through Experience

Explore artificial intelligence and large language models through beautiful, interactive demonstrations. Learn by doing, not just reading.

Start LearningBrowse Topics
Topics
68
Interactive guides to understand AI concepts
Recently Updated
Mega-Kernels
Sep 24, 2026
LLM
Architecture
MTP · MoE · KV cache · inference

LMF Blog

Latest AI news, model analysis, and deep dives

Interactive Demos

Visual Learning

Build Intuition

Recently Updated

What's New

ESep 24, 2026
Mega-KernelsFuse operations and move whole transformer steps into a single GPU kernel to cut launch overhead and HBM traffic
BSep 13, 2026
Getting StartedMake your first LLM API call in 10 minutes — for free
BSep 13, 2026
The Agent LoopThe observe-think-act cycle that powers autonomous agents
ISep 13, 2026
Context AnatomyHow agents structure and manage their context window
ISep 13, 2026
Tool DesignPrinciples for building tools that agents can use reliably

Help Make This Better

This guide is open source. Got an idea for a new topic? Found a bug? Want to improve an explanation? Every contribution helps.

Request a TopicReport a BugStar on GitHub

Explore Topics

Dive into interactive lessons

Artificial Intelligence

01
Getting StartedMake your first LLM API call in 10 minutes — for freeSep 13, 2026
B
AI Agents
Large Language Models
Architecture
01
Transformer ArchitectureThe foundational architecture behind GPT, BERT, and all modern LLMsSep 13, 2026
I
02
Looped TransformersExplore how recurrent depth updates a hidden state, reuses model weights, and changes the tradeoffs between memory, compute, latency, and monitorability.Sep 7, 2026
E
03
Feed-Forward NetworksHow the same MLP transforms each token position; routing is covered in Mixture of Experts.Sep 13, 2026
I
04
Residual Stream & LayerNormHow residual streams and normalization keep deep transformer stacks trainable.Sep 13, 2026
I
05
Next-Token PredictionHow logits, softmax, and decoding turn model outputs into text one token at a time.Sep 13, 2026
B
06
LLM TrainingFrom pretraining on raw text to fine-tuning with human feedbackSep 13, 2026
I
07
Training DataWhere AI models get their knowledge: legitimate sources, controversies, and synthetic dataSep 13, 2026
I
08
Mixture of ExpertsRun only a fraction of model parameters per tokenSep 13, 2026
E
09
QuantizationShrink model size by reducing numerical precisionSep 13, 2026
E
10
Nested LearningLearning algorithms that operate at multiple levelsSep 13, 2026
E
11
Multi-Token Prediction (MTP)Train models to predict several future tokens at once instead of only the next tokenSep 13, 2026
E
12
N-gram EmbeddingsLearned lookup tables for short local token patternsAug 27, 2026
E
13
DistillationTransfer knowledge from a large model to a smaller oneSep 13, 2026
I
14
Fine-Tuning & LoRAAdapt large models efficiently by training only tiny low-rank matricesSep 13, 2026
I
15
AbliterationRemove a model's refusal behavior by ablating a single direction in its residual streamSep 13, 2026
E
16
Speculative DecodingSpeed up inference by drafting tokens with a smaller modelSep 13, 2026
E
Diffusion Models
01
How Diffusion WorksForward noise, reverse denoising, and why score estimation worksSep 13, 2026
I
02
Text DiffusionMask-and-predict generation with fixed-length token gridsSep 13, 2026
I
03
Image DiffusionLatent denoising pipelines behind modern text-to-image systemsSep 13, 2026
I
01
KV CacheStore computed keys and values to avoid redundant workSep 13, 2026
E
02
Mega-KernelsFuse operations and move whole transformer steps into a single GPU kernel to cut launch overhead and HBM trafficSep 24, 2026
E
03
Prompt CachingReuse computed KV caches across API requests to save cost and latencySep 13, 2026
I
04
Batching & ThroughputProcess multiple requests simultaneously for higher throughputSep 13, 2026
I
05
Running Models LocallyRun LLMs on your own hardware for privacy, speed, and zero API costsSep 13, 2026
B
06
VRAM CalculatorCalculate a local LLM memory budget from explicit weight, cache and runtime assumptionsSep 13, 2026
B
01
Neural NetworksLayers of connected neurons that learn patterns from dataFeb 9, 2026
B
02
Gradient DescentThe optimization algorithm that trains neural networksSep 13, 2026
I
03
Training ProcessHow models learn from data through forward and backward passesSep 13, 2026
B
04
Reinforcement LearningAgents learn by acting, receiving rewards, and improving a policySep 13, 2026
I
05
World ModelsInternal simulators that let AI predict and plan in virtual worldsSep 13, 2026
E
Prompting
01
Prompt BasicsFundamental techniques for getting good results from LLMsSep 13, 2026
B
02
Advanced TechniquesChain-of-thought, few-shot, and other power techniquesSep 13, 2026
I
03
System PromptsSet the persona, rules, and behavior of an LLMSep 13, 2026
I
AI Safety & Ethics
01
Bias & FairnessHow training data biases surface in model outputsSep 13, 2026
B
02
Responsible AIBuilding AI systems that are fair, transparent, and safeSep 13, 2026
B
AI Industry
01
AI Made in EuropeThe growing European AI ecosystem and its unique strengthsSep 13, 2026
B
02
Open Source AdvantagesWhy open weights matter for innovation and transparencySep 13, 2026
B
03
Custom Chips for AIWhy GPUs, TPUs, NPUs, FPGAs, and AI ASICs shape model cost and speedSep 13, 2026
I
04
Logge’s Model NotesGPT-6 Astra as the main driver, a current cost frontier and earlier personal rankings.Sep 13, 2026
B
Pro tip: PressCtrl+Kto search topics