The Annotated JEPA | Elements of a Vector Space

An annotated walkthrough of Joint Embedding Predictive Architectures.

--> This post is a step-by-step, annotated, from-scratch walkthrough of Joint Embedding Predictive Architectures, or JEPAs. The goal is to do for JEPA what The Annotated Transformer did for the Transformer: build the full object, explain every moving part, and end with a working training loop. JEPA is Yann LeCun's proposed answer to a fundamental question in self-supervised learning: how do you train a model to understand the world without labels, without collapsing to trivial solutions, and without wasting capacity on irrelevant details? The answer, elegant in principle and subtle in practice

saved by

• Akira Yoshiyama
• Jaisel Singh
• Ishaan Panigrahi
• Omkar Gadewar
• Aarya Gadekar

related reading

• The first AI model based on Yann LeCun’s vision for more human-like AIai.facebook.com
• the j-lens: finding an llm's unspoken concepts / chiragctxnn.github.io
• V-JEPA: The next step toward advanced machine intelligenceai.meta.com
• LeWorldModel: Stable End-to-End Joint-Embedding Predictive Architecture from Pixelsle-wm.github.io
• [2509.14252] LLM-JEPA: Large Language Models Meet Joint Embedding Predictive Architecturesarxiv.org
• [2506.09985] V-JEPA 2: Self-Supervised Video Models Enable Understanding, Prediction and Planningarxiv.org
• arxiv.org/pdf/2511.08544arxiv.org
• [2603.14482] V-JEPA 2.1: Unlocking Dense Features in Video Self-Supervised Learningarxiv.org
• Stanford CS25: Transformers United V6 I Overview of Transformersyoutube.com
• Stanford CS25: Transformers United V6 I From Representation Learning to World Modelingyoutube.com
• [2602.10098] VLA-JEPA: Enhancing Vision-Language-Action Model with Latent World Modelarxiv.org
• Causal-JEPA: Learning World Models through Object-Level Latent Maskingarxiv.org