# dabochen/spreadsheet-is-all-you-need

A nanoGPT pipeline packed in a spreadsheet

Repository: https://github.com/dabochen/spreadsheet-is-all-you-need
Canonical: https://ross.abutalabs.com/products/spreadsheet-is-all-you-need
License: MIT
License Family: permissive
Last push: 2024-06-17T00:08:59+00:00

## Health v2 (maintenance only)
Score: 24/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 58
- inputs: {"age_days": 816, "days_push": 808, "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 2159, forks 128 (observed 2026-08-28T04:06:20.747134+00:00)

## What it is
An interactive spreadsheet (Excel/Numbers) that implements the full nanoGPT inference pipeline, including embedding, layer norm, self-attention, MLP, and softmax, with roughly 85,000 parameters. It is a visual educational tool for understanding transformer internals through fully configurable matrix calculations.

## Use cases
- visualize how a transformer processes data step by step
- understand self-attention with interactive matrix calculations
- learn GPT architecture without writing Python code
- experiment with transformer parameters in a spreadsheet
- teach deep learning concepts with a hands-on tool
- explore nanoGPT's inference pipeline interactively

## When to choose
- you are a visual learner trying to understand transformer internals
- you want an interactive, configurable walkthrough of GPT mechanisms
- you want to teach or self-study how self-attention and MLPs work
- you prefer spreadsheets over code for exploring calculations

## When to avoid
- you need actual trained weights or correct inference outputs
- you want to train or fine-tune a language model
- you need production-grade LLM inference performance
- you expect large-scale tokenization beyond the A/B/C character demo

## Facets
- artifact type: learning-resource
- maturity: stable
- function: machine-learning, data-visualization
- domain: deep-learning, large-language-models, education, tutorials
- platform: windows, cross-platform
- tags: nanogpt, transformer, spreadsheet, self-attention, interactive-learning, llm-internals, education, macos

## Member repositories
- dabochen/spreadsheet-is-all-you-need (main) score 24

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:20.747134+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 2026-08-30T02:50:19.689118+00:00, confidence not recorded.
  - readme: https://github.com/dabochen/spreadsheet-is-all-you-need (fetched 2026-08-28T04:06:20.747134+00:00, sha 8662ccae46c8)
- Data as of 2026-08-30T08:39:29.467469+00:00.