openai / automated-interpretability
Components

An overview of source code logical components.

Intro

Logical decomposition is a representation of the organization of the main source code, where every and each file is put in exactly one logical component.

Logical Decompositions Overview

Analyzed system has 1 logical decomposition:

Logical Decomposition #1: PRIMARY

The decompositions is based on the folder structure at level 1 (relative to the source code root).

Bubble Chart | Tree Map
Component Sizes (Lines of Code)
The "primary" logical decomposition has 2 components.
  • 41 files, 4,292 LOC (100.0% vs. main code).
  • "neuron-explainer" is biggest, containing 68.83% of LOC.
  • "neuron-viewer" is smallest, containing 31.17% of LOC.


neuron-explainer2954 LOC (68%) 20 files
neuron-viewer1338 LOC (31%) 21 files
Component Commits
Components ordered by number of commits
Total Commits per Component
All commits, some commits may include files from multiple components.
neuron-explainer18 commits (69%)
neuron-viewer3 commits (11%)
Yearly File Updates Trend per Components
The number of file changes in commits
animated commit history: all time cumulative | 12 months window
2025 2024 2023
neuron-explainer
5
17
neuron-viewer
3


Dependencies between components in same commits (past 180 days)
The number on the lines shows the number of shared commits.
See detailed temporal dependencies report...

No temporal dependencies found.



2025-05-04 14:11