IIT Expert

Precomputed result

The five-unit network of the IIT 4.0 paper in state ABcdE specifies 23 distinctions and 13740 relations

Substrate Five binary units A–E with states in {−1, 1}, each a logistic of its weighted inputs (Eq 60) with k = 4. Strong connections of weight 0.8 run around the ring A→B→C→D→E→A. Inhibitory connections of weight −0.05 run the other way (B→A, C→B, D→C, E→D, A→E). D→A has weight 0.4, and all other connections, including the self-connections, have weight 0.05. PyPhi ships the substrate, with the weights supplied by the authors, as pyphi.examples.iit4_2023_fig7_substrate.
State (1, 1, −1, −1, 1) — ABcdE, units A, B, E ON
φs 0.477459747138178
Minimum partition A directed partition severing 14 of the 25 connections; the script lists them. The same under both formulations.
Formalism iit-4.0-2026
PyPhi 2.0.0rc4
Script albantakis-2023-fig7a-active.py

Albantakis et al. (2023) use this network to show how the state of a single unit changes the Φ-structure. Panel A is the reference state, with unit E active. Panels B and C change only unit E: in panel B it is OFF, and in panel C it is inactivated (iit:result/albantakis-2023-fig7c-inactivated).

quantity Fig 7A 2023 2026
complex ABcdE ABcdE ABcdE
φs 1.1 1.1056 0.4775
distinctions 23 of 31 23 23
relations 13740 13740 13740
Φ 22.26 22.2665 22.2665

The 23 distinctions match the figure's list in mechanism, cause purview and state, effect purview and state, and φd to four decimals (for example A → E | B with φd = 0.7837 and ABcdE → cDE | BC with 0.0028). Σφd = 4.5027 and Σφr = 17.7638. The figure truncates Φ to 22.26.

The Φ-structure is the same under both formulations. The intrinsic information requirement of Mayner et al. (2026) lowers φs from 1.11 to 0.48; the full system remains the complex.

The script uses the default configuration of each formulation. Relations are summed analytically; their number follows from the distinctions. To reproduce, run the script with uv, which installs PyPhi 2.0.0rc4 for the run:

uv run --with "pyphi==2.0.0rc4" python canon/albantakis-2023-fig7a-active.py

Sources