← Python API
vraven.darkside · function

build_circuit_wiring

build_circuit_wiring(model: 'nn.Module', inputs: 'Any', circuitry: 'AnalysisResult', *, include_removed: 'bool' = False, minimum_edge_mass: 'float' = 0.0) -> 'AnalysisResult'

Build a neuron-group wiring graph from an ε-sufficient circuit core.

Nodes are retained (optionally removed) unit groups from
``discover_decision_circuitry``. Edges between consecutive Linear layers
use the measured mass:

mass(G → H) = Σ_{i∈G, j∈H} |W_{j i}| · |a_i|

where ``a`` is the captured input activation to the downstream Linear layer
and ``W`` is that layer's weight matrix. This is not a claim of unique
causal edges; it is the measured weight×activation coupling of the core.

Usage

from vraven.darkside import build_circuit_wiring

result = build_circuit_wiring(... )
Source of truthThis page is generated from the checked-out public symbol, signature, and docstring. Pin the package version for reproducible work.