# openai/automated-interpretability

Repository: https://github.com/openai/automated-interpretability
Canonical: https://ross.abutalabs.com/products/automated-interpretability
Language: Python
License Family: other
Archived: true
Last push: 2024-03-06T18:40:46+00:00

## Health v2 (maintenance only)
Score: 10/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 86
- inputs: {"age_days": 1213, "days_push": 910, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, archived, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1081, forks 130 (observed 2026-08-28T04:03:30.797007+00:00)

## What it is
OpenAI's code and tools for automatically generating, simulating, and scoring explanations of neuron behavior in language models, based on the 'Language models can explain neurons' paper. It also includes a neuron activation viewer and public datasets of GPT-2 XL neuron activations and explanations.

## Use cases
- automatically explain what individual neurons in a language model do
- generate and score explanations of neuron behavior using LLMs
- visualize neuron activations and their explanations
- analyze GPT-2 XL neuron activation datasets
- research mechanistic interpretability of transformer models
- simulate neuron behavior to validate generated explanations

## When to choose
- you want to interpret or explain neuron behavior inside large language models
- you need OpenAI's official tooling for neuron-level interpretability research
- you want to explore the released GPT-2 XL neuron activation and explanation datasets
- you are replicating or extending the neuron-explanation methodology from the paper

## When to avoid
- you need a production-ready, actively maintained library with support guarantees
- you want general-purpose model explainability for classifiers rather than LLM neurons
- you cannot access Azure blob storage for the public datasets
- you need interpretability tooling for non-Python stacks

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, nlp, llm-inference, data-visualization, developer-tools
- domain: machine-learning, deep-learning, large-language-models, artificial-intelligence, developer-tools
- platform: python
- tags: interpretability, neuron-explanations, mechanistic-interpretability, gpt-2, explainability, research

## Member repositories
- openai/automated-interpretability (main) score 10

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:30.797007+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 2026-08-30T06:51:23.406554+00:00, confidence not recorded.
  - readme: https://github.com/openai/automated-interpretability (fetched 2026-08-28T04:03:30.797007+00:00, sha 85e4ca7cc757)
- Data as of 2026-08-30T08:39:29.467469+00:00.