Legacy Concept Lab

Probing, Linear Classifier Probes & Activation Analysis

Probing is one of the main tools to understand what GPT-like models know and where that knowledge lives

Concept 12 of 100RepresentationsPhase 5
#12ProbingRepresentations
key equation
y^=Wh(x)+b\hat y = W h_\ell(x) + b
Phase 5: Representation & interpretabilityConcept 12 of 100

Why It Matters for Modern Models

  • Probing is one of the main tools to understand what GPT-like models know and where that knowledge lives
  • Used heavily for safety (probing for dangerous capabilities), robustness, and fairness analyses

What Tutorials Skip

What is still poorly explained in textbooks and papers:

  • Clear mental model of what probes measure (information content vs ease of extraction)
  • Visual, layer-by-layer maps of information flow in large LMs

Interactive Visualization

Core Math (Optional Deep Dive)

If you want intuition first, start with the key equation and the visualization. Come back here for the full walkthrough.

Key Equation
y^=Wh(x)+b\hat y = W h_\ell(x) + b

Given layer representation h(x)h_\ell(x), train a frozen probe:

y^=Wh(x)+b\hat y = W h_\ell(x) + b

(or a softmax over Wh(x)Wh_\ell(x)) on a supervised task (POS tags, parse trees, etc.). The accuracy estimates how linearly separable that information is at layer \ell.

BERT layers roughly follow the classical NLP pipeline (POS → syntax → semantics → coreference).

Canonical Papers

Understanding Intermediate Layers using Linear Classifier Probes

Alain & Bengio2016ICLR Workshop
Read paper →

BERT Rediscovers the Classical NLP Pipeline

Tenney et al.2019ACL
Read paper →

Connections

Next Moves

Explore this concept from different angles — like a mathematician would.