Topics

Curriculum Map

Browse every block and topic in the recommended sequence.

1. 1

   Transformer Foundations

   8 topics

   • Prerequisites
   • Transformer Architecture Intro
   • The Attention Mechanism
   • MLPs in Transformers
   • Layer Normalization
   • QK and OV Circuits
   • Composition and Virtual Attention Heads
   • Decoding Strategies
2. 2

   Interpretability Fundamentals

   3 topics

   • What is Interpretability?
   • The Linear Representation Hypothesis
   • The Superposition Hypothesis
3. 3

   Basic Techniques

   3 topics

   • The Logit Lens and Tuned Lens
   • Direct Logit Attribution
   • Reading the Attention Patterns
4. 4

   Causal Interventions

   5 topics

   • Activation Patching and Causal Interventions
   • Attribution Patching and Path Patching
   • Refined Attribution Methods
   • Self-Repair in Language Models
   • The Causal Abstraction Framework
5. 5

   Probing

   6 topics

   • Probing Classifiers
   • Truthfulness Probing and the Geometry of Truth
   • Contrastive Activation Addition (CAA)
   • Linear Artificial Tomography (LAT)
   • Attention Probes
   • Probes in Production
6. 6

   Steering

   7 topics

   • Addition Steering
   • Ablation Steering
   • Affine Steering
   • Representation Control
   • Function Vectors
   • Unsupervised Steering Vectors
   • Multi-Layer Steering
7. 7

   Model Editing

   2 topics

   • Localized Fact Editing and Its Pitfalls
   • Concept Erasure with LEACE
8. 8

   Superposition & Feature Extraction

   7 topics

   • Sparse Autoencoders: Decomposing Superposition
   • Feature Dashboards and Automated Interpretability
   • Scaling Monosemanticity and Feature Steering
   • SAE Variants, Evaluation, and Honest Limitations
   • Feature Geometry: Beyond One-Dimensional Directions
   • Transcoders: Interpretable MLP Replacements
   • Crosscoders
9. 9

   Model Diffing

   3 topics

   • Logit Diff Amplification
   • Feature-Level Model Diffing
   • Finetuning Traces in Activations
10. 10

    Hidden State Decoding

    6 topics

    • Hidden State Decoding: From Vectors to Language
    • Patchscopes
    • SelfIE: Self-Interpretation of Embeddings
    • Training Models to Explain Their Computations
    • LatentQA and Latent Interpretation Tuning
    • Activation Oracles
11. 11

    Circuit Finding

    7 topics

    • Induction Heads and In-Context Learning
    • The IOI Circuit: Discovery and Mechanism
    • Circuit Evaluation: Faithfulness, Completeness, and Minimality
    • Circuit Tracing and Attribution Graphs
    • Universality Across Models
    • Multimodal Mechanistic Interpretability
    • Copy Suppression
12. 12

    Black-Box Interpretability

    1 topics

    • Counterfactual Resampling
13. 13

    MI for AI Safety

    5 topics

    • The Refusal Direction
    • Detecting Sleeper Agents with Mechanistic Interpretability
    • Deception Detection and Alignment Faking
    • Understanding Safety Mechanisms and MI-Based Monitoring
    • Honest Limitations of MI for Safety
14. 14

    Tools

    3 topics

    • TransformerLens
    • nnsight and nnterp
    • SAELens and Neuronpedia
15. 15

    More Resources

    2 topics

    • ARENA: Hands-On Technical Training
    • Getting Started in MI Research