Every premium and members-only resource for $9.99/month, as a no-commitment subscription. See the plans

PalateKit

Sign in
Menu

Request a resource

Describe the resource you are missing in one sentence. We will email you when it is published.

0 / 280

A page, a video or an image online.

Unlock PalateKit

Unlock everything

Plan

Paying means accepting our terms of sale.

Secure payment by Stripe. Receipts and invoices sent by email.

Welcome back

Sign in to install, download or save this resource. A free account is enough for free resources.

Mechanistic interpretability with TransformerLens v1.0.0

Use this resource to study the internal mechanisms of a transformer through activation patching and attention analysis

New

  • 527 downloads
  • 155 likes
  • 144 saves

Free

Install with the CLI

In a terminal, from the folder you want:

npx --yes palatekit@latest install transformer-internals-interpretability

With no option, the skill goes to ~/.agents/skills/; add --agent claude-code (or codex, cursor…) to target an agent.

First time: connect the CLI with

npx --yes palatekit@latest login

The agent writes the code that loads a model into TransformerLens, caches its activations and runs patching experiments. You get an analysis script and the right APIs to inspect attention patterns.

Mostly picked by

Updated November 2, 2026

Card for the “Mechanistic interpretability with TransformerLens” skill

Details

Works in
Install guides
Generates with
Your agent's model
Licence
MIT

Similar resources

View all

What you get

  • SKILL.md: concepts (HookedTransformer, activation cache) and installation
  • references/api.md: API reference
  • references/tutorials.md: step-by-step tutorials

Good fit for

  • You are looking for the algorithm a model learned during training
  • You run activation patching experiments on open models

Not for

    Tags

    • interpretability
    • transformers
    • activations
    • research

    About

    You say

    “Cache GPT-2 small's activations and compare the attention patterns of layer 5 heads.”

    Card for the “Mechanistic interpretability with TransformerLens” skill

    Skill published by Orchestra Research under the MIT license, delivered in its original version.

    Show more

    How to use it

    1. Install the skill in your agent's skills folder
    2. Ask your agent to study the internal mechanisms of a transformer through activation patching and attention analysis
    3. Review and adjust the result

    Content preview

    Files

    • LICENSE1.1 KB
    • NOTICE-PALATEKIT.md0.5 KB
    • references/api.md8.5 KB
    • references/README.md1.7 KB
    • references/tutorials.md10.2 KB
    • SKILL.md12.1 KB

    Origin and licence

    Listed unmodified

    Author
    Orchestra Research
    Source
    github.com/orchestra-research/ai-research-skills/tree/773a52944ba4747a18bd4ae9ade53fff041adcbc/04-mechanistic-interpretability/transformer-lens
    Licence
    MIT

    Notice

    « transformer-lens-interpretability » (Orchestra Research, MIT) livré dans sa version d'origine : fiche en français et en anglais par PalateKit.

    User reviews

    –

    No ratings yet

    Loading reviews…

    Delete your review?

    Your rating and comment will be removed from the page.