Right in your browser
Understand the concepts
Look up terms such as tokenizer and attention in the Web Wiki. Use Explain to explore highlighted terms right here – no installation needed.
For Windows · Desktop applicationUnderstand language models.
UNDERSTAND LANGUAGE MODELS · STEP BY STEP
Understand language models.
By building one yourself.
M.E.F LLM Studio is an educational app for beginners: understand language models through clear explanations and small, hands-on experiments, without writing scripts. Follow the path from tokens through training to generated text. Brain Explorer, Wiki and Vision also make text and image generation explorable step by step.
CPU starting paths are available: chapters, data preparation and supported pretrained models do not require an NVIDIA GPU. Computation can take substantially longer on CPU and needs enough RAM. Training and Vision have their own requirements; the guide explains the limits for each path. Explore the CPU starting path →
Where would you like to start?
In the Windows app
Experiment for yourself
Go from your own text through a tokenizer and dataset to a small language model. The step-by-step guide walks you through the Windows app.
Browser demo and app
Look inside the model
Follow a recorded text generation in Demo Explorer, from context to the next token. In the app, you can also explore how images are generated.
Why I am building this
How do I explain LLMs to my family?
The more I worked with LLMs, the more I wanted to understand what happens behind the answers. And how could I explain it to my son or my wife, who use these models without knowing how they work?
That is where M.E.F LLM Studio came from: trying things yourself, making connections visible and taking away a little of the unease around the apparent magic.
Read the whole story →M.E.F LLM STUDIO
From an explanation to your own experiment.
Read what a tokenizer does, then inspect your own text. Learn what loss means, then follow a training run. The learning overview connects the fundamentals with practical tools, turning terminology into relationships you can observe.
The Windows application’s learning overview. The progress shown is an example state.
Getting started ↗
DATA
How does text become numbers?
Split your own sentences into tokens and explore why tokens are not simply words.
RUNS
What does training change?
Follow loss and validation and explore what these measurements tell you about a run—and what they do not.
MODELS
Why keep checkpoints?
Save model states and use them to compare results at different points in training.
CHAT
How does the next reply emerge?
Try a prompt and inspect how sampling affects the selection of the next tokens.
01 — THE LEARNING PATH
Read. Experiment. Make sense of it.
Repeat this learning cycle for each topic. Small, manageable experiments are the focus; you do not need to train a large model.
Understand a question
Read the relevant chapter: What are tokens? How does a model learn? What does a loss curve tell you?
Run your own experiment
Work with a short text or a small model. Change one setting deliberately and observe what happens.
Interpret the result
Compare your observation with the explanation. Even nonsensical output is a reason to examine the data, training and limitations more closely.
02 — M.E.F LLM STUDIO
Understand what your model does.
Go beyond training a model. Explore how it processes text and compare its predictions across saved training checkpoints.
Explore learning and comparison features ↗New in 0.2.0
From a concept to a visible computation.
Alpha 0.2.0 extends the learning path with Brain Explorer, Wiki and Vision: understand concepts, run your own experiments and inspect the computation step by step.
Brain Explorer
How is a reply built, token by token?
A 2D view connects context, transformer blocks, probabilities and sampling. Play a recording, pause it or inspect each step individually.
Wiki & Explain
What does the term on this screen mean?
Learning chapters and a searchable Wiki connect fundamentals to experiments. Explain mode describes highlighted terms beside the visible interface.
Vision & Vision Explorer
How does noise become an image?
Generate images locally with a supported model and open the recording directly in the Vision Explorer. Follow denoising step by step: state before, noise prediction and state after.