# bbycroft/llm-viz

3D Visualization of an GPT-style LLM

Repository: https://github.com/bbycroft/llm-viz
Canonical: https://ross.abutalabs.com/products/llm-viz
Homepage: https://bbycroft.net
Language: TypeScript
License: MIT
License Family: permissive
Last push: 2026-08-10T22:12:08+00:00

## Health v2 (maintenance only)
Score: 74/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 97, release rhythm 35, longevity 91
- inputs: {"age_days": 1281, "days_push": 23, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 5509, forks 658 (observed 2026-08-28T04:09:21.781975+00:00)

## What it is
An interactive 3D visualization and walkthrough of a GPT-style LLM running inference, showing the network down to individual add and multiply operations. It renders the transformer topology used in GPT-2/GPT-3 and includes a working tiny example model from Karpathy's minGPT.

## Use cases
- visualize how a GPT transformer works step by step
- learn LLM inference internals interactively
- understand attention and transformer architecture in 3D
- teach students how ChatGPT's underlying algorithm works
- explore a working GPT model down to individual operations

## When to choose
- you want an intuitive, visual explanation of transformer inference
- you're teaching or learning how GPT-style networks compute
- you prefer interactive 3D walkthroughs over papers or blog posts

## When to avoid
- you need a production LLM training or inference tool
- you want to visualize your own model's weights at scale (large weight downloads unsupported)
- you need a general-purpose neural network debugger

## Facets
- artifact type: learning-resource
- maturity: active
- function: data-visualization, machine-learning, graphics
- domain: large-language-models, education, deep-learning, data-visualization
- platform: browser
- tags: 3d-visualization, gpt, transformer, interactive-walkthrough, inference, webgl, web, typescript

## Member repositories
- bbycroft/llm-viz (main) score 74

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:09:21.781975+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T17:56:15.832346+00:00, confidence not recorded.
  - readme: https://github.com/bbycroft/llm-viz (fetched 2026-08-28T04:09:21.781975+00:00, sha c9917bdc7ff8)
  - homepage: https://bbycroft.net (fetched 2026-08-29T08:51:31.680794+00:00, sha 39abe0b0a0fb)
- Data as of 2026-08-30T08:39:29.467469+00:00.