---
title: JEPA Wiki
emoji: 🧠
colorFrom: gray
colorTo: gray
sdk: docker
app_port: 3000
pinned: false
---

# JEPA Wiki

A Wikipedia-style knowledge base tracking the evolution of **Joint-Embedding Predictive Architecture (JEPA)** research — from I-JEPA (2023) through V-JEPA 2.1, ThinkJEPA, and beyond.

**19 papers** covered across the full JEPA family plus related work (DINO, DINOv2, NEPA), with **8 concept pages**, **15 animated video explainers** (manim), and **7 SVG architecture diagrams**.

Built with Next.js, rendered from markdown wiki files with `[[wikilink]]` support.

## How this was built

This wiki was built entirely through conversation with [Claude Code](https://claude.ai/claude-code), using the [LLM Wiki](https://gist.github.com/tobi/1115fc2e0537ea47aba18ce3e773e3a8) pattern by Tobi Lutke (inspired by Karpathy's endorsement of the approach).

The workflow:

1. **Source discovery**: started from [Turing Post: 14 JEPA Milestones as a Map of AI Progress](https://www.turingpost.com/p/jepamap) article, extracted all arxiv paper IDs
2. **Paper ingestion**: used [`hf papers read`](https://huggingface.co/docs/huggingface_hub/en/guides/cli#huggingface-cli-papers) (Hugging Face CLI) as the primary tool to fetch full paper markdown from arxiv HTML — the agent reads papers, extracts key concepts, and writes wiki pages
3. **Incremental expansion**: discovered 7 additional papers (V-JEPA, LeJEPA, IWM, LLM-JEPA, VL-JEPA, EB-JEPA) by searching for related work by Yann LeCun and Saining Xie
4. **Cross-referencing**: concept pages (masking strategies, collapse prevention, world models, etc.) synthesize information across all papers with concrete numbers and comparisons
5. **Visualization**: 21 short (5-12s) manim animations created using the [Hermes Agent manim-video skill](https://github.com/NousResearch/hermes-agent/tree/main/skills/creative/manim-video) — a skill that provides structured guidelines for creating 3Blue1Brown-style educational animations with Manim Community Edition. The skill's visual design standards (opacity layering, monospace fonts, dark backgrounds, one-idea-per-scene) shaped the animation style. SVG architecture diagrams for static concepts.

## Resources

- [Turing Post: 14 JEPA Milestones](https://www.turingpost.com/p/jepamap) — the article that seeded this wiki
- [LLM Wiki pattern](https://gist.github.com/tobi/1115fc2e0537ea47aba18ce3e773e3a8) — Tobi Lutke's idea file for building personal knowledge bases with LLMs
- [`hf papers`](https://huggingface.co/docs/huggingface_hub/en/guides/cli#huggingface-cli-papers) — HF CLI tool for reading arxiv papers as markdown (`hf papers read <arxiv_id>`)
- [Hermes Agent manim-video skill](https://github.com/NousResearch/hermes-agent/tree/main/skills/creative/manim-video) — Nous Research skill for creating educational manim animations, used to generate the 21 video explainers
- [LeCun: A Path Towards Autonomous Machine Intelligence](https://openreview.net/pdf?id=BZ5a1r-kVsf) — the foundational JEPA position paper