OV Circuit

The circuit formed by W_U W_OV^h W_E that describes how a given token affects output logits if attended to

Neighborhood — ranked by edge-count

paper

method

Freezing Attention Patterns Trick
about
A conceptual technique of fixing attention patterns to make the transformer a purely linear function of tokens, enabling independent analysis of OV and QK circuits

concept

Copying Matrix
associated_with
An OV circuit matrix that maps tokens to increasing the logit of those same tokens; detectable via positive eigenvalues

cosine ≥ 0.65 · no typed edge

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

Equivariant Circuitconcept0.723
A circuit exhibiting symmetry such that the weights rotate with the orientation of the feature they detect
Circuit Findingmethod0.713
Interpretability technique for identifying functional sub-circuits in neural networks, supported by pyvene
Circuits Threadframework0.704
An open scientific collaboration hosted on Distill slack studying the inner workings of neural networks via zoomed-in mechanistic investigation
QK Circuitconcept0.703
The bilinear circuit formed by W_E^T W_QK^h W_E that determines which source token a destination token attends to
Recurrent Logic Circuitconcept0.701
The key novel property of DiffLogic CA — logic gate networks that are recurrent both spatially and temporally
Circuit Motifconcept0.696
A recurring, abstract pattern found in circuits (e.g., equivariance, unioning over cases), inspired by circuit motifs in systems biology
Circuit Interpretabilityconcept0.694
Advantage of DiffLogic CA over NCA — learned rules are pure binary logic circuits that can be visualized and analyzed
Minimal Circuit Descriptionconcept0.693
Hypothesis that discretization finds minimum-size circuits equivalent to minimal algorithmic descriptions of patterns