pyphi.conf.presets#

Canonical IIT formalism presets.

Each preset is a dict consumable by config.override():

  • iit3 — IIT 3.0 (Oizumi et al. 2014).

  • iit4_2023 — IIT 4.0 (Albantakis et al. 2023).

  • iit4_2026 — IIT 4.0 with the intrinsic-information requirement (Mayner, Marshall, Tononi 2026).

Usage:

from pyphi import config, iit3

with config.override(**iit3):
    ...  # computations use the IIT 3.0 formalism

A preset specifies its formalism sub-namespaces wholesale via IITConfig and ActualCausationConfig instances, so applying a preset resets every field in those sub-namespaces to its canonical value for that paper. Fields outside the formalism layer (e.g. precision) appear as additional top-level keys when the paper’s behavior depends on them.

Functions

canonical(formalism)

Return the version name that formalism selects.