{
  "slug": "transformer-explainer",
  "url": "/instruments/transformer-explainer/",
  "title": "Transformer Explainer: GPT-2 Live in the Browser",
  "type": "instrument",
  "typeLabel": "Instrument · adaptation",
  "lifecycle": "instrument",
  "epistemicStatus": "adaptation",
  "statusLabel": "Adapted instrument · original by Cho et al., Georgia Tech Polo Club",
  "programme": "computation",
  "question": "What does a transformer's forward pass actually compute, step by step, on real input?",
  "date": "2026-07-10",
  "version": "v1",
  "authors": [
    "Latent Minds Institute"
  ],
  "methods": [
    "ONNX GPT-2 (small) inference in-browser",
    "attention map inspection",
    "temperature and sampling controls"
  ],
  "models": [
    "GPT-2 small (124M)"
  ],
  "evidence": "pedagogical instrument running a real model; the 2026 commentary is editorial, not original research",
  "code": "https://github.com/poloclub/transformer-explainer",
  "data": null,
  "provenance": "Adaptation of poloclub/transformer-explainer (MIT); previously hosted as Transformer Visualiser at mo3.ca",
  "related": [
    "interpretability-map"
  ],
  "canonicalUrl": "https://latentmindsinstitute.com/instruments/transformer-explainer/",
  "apiUrl": "https://latentmindsinstitute.com/api/research-objects/transformer-explainer.json"
}