Topics

Curriculum Map

Browse every block and topic in the recommended sequence.

1. 1

   Transformer Foundations

   10 topics

   • Prerequisites
   • Transformer Architecture Intro
   • Embeddings
   • Positional Embeddings
   • 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 Mechanistic Interpretability?
   • The Linear Representation Hypothesis
   • The Superposition Hypothesis
3. 3

   Basic Techniques

   4 topics

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

   Causal Interventions

   6 topics

   • Activation Patching and Causal Interventions
   • Choosing Causal Mediators
   • 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

   8 topics

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

   Model Editing

   3 topics

   • Localized Fact Editing and Its Pitfalls
   • LEACE and Linear Concept Erasure
   • Memorization and Machine Unlearning
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 Limitations
   • Temporal Representations and Feature Extraction
   • Transcoders: Interpretable MLP Replacements
   • Crosscoders
9. 9

   Feature Geometry

   3 topics

   • Feature Geometry: Beyond One-Dimensional Directions
   • Discovering and Interpreting Neural Manifolds
   • Manifold Steering
10. 10

    Model Diffing

    3 topics

    • Logit Diff Amplification
    • Feature-Level Model Diffing
    • Finetuning Traces in Activations
11. 11

    Hidden State Decoding

    8 topics

    • Hidden State Decoding: From Vectors to Language
    • Patchscopes
    • SelfIE: Self-Interpretation of Embeddings
    • Testing Introspection with Concept Injection
    • Training Models to Explain Their Computations
    • LatentQA and Latent Interpretation Tuning
    • Activation Oracles
    • Natural Language Autoencoders
12. 12

    Circuit Finding

    8 topics

    • Induction Heads and In-Context Learning
    • The IOI Circuit: Discovery and Mechanism
    • Circuit Evaluation: Faithfulness, Completeness, and Minimality
    • Entity Binding and Retrieval
    • Circuit Tracing and Attribution Graphs
    • Universality Across Models
    • Multimodal Mechanistic Interpretability
    • Copy Suppression
13. 13

    Black-Box Interpretability

    1 topics

    • Counterfactual Resampling
14. 14

    Weight-Space Interpretability

    3 topics

    • Parameter Decomposition
    • Circuits in Parameter Space
    • Training Models to Be Interpretable
15. 15

    MI for AI Safety

    6 topics

    • The Refusal Direction
    • Detecting Sleeper Agents with Mechanistic Interpretability
    • Deception Detection and Alignment Faking
    • Evaluation Awareness and Evaluation Validity
    • Understanding Safety Mechanisms and MI-Based Monitoring
    • Limits of Mechanistic Interpretability for Safety
16. 16

    Tools

    3 topics

    • TransformerLens
    • nnsight and nnterp
    • SAE Lens and Neuronpedia
17. 17

    More Resources

    2 topics

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