IIT 3.0

Version 3.0 of IIT, used to compute Φmax and Conceptual Information (CI) from LLM representation networks.

Neighborhood — ranked by edge-count

thinker

Masafumi Oizumi
introduces
First author of Oizumi, Albantakis, & Tononi (2014) 'From the phenomenology to the mechanisms of consciousness: integrated information theory 3.0'.

method

PyPhi
implements
Software toolkit used to compute Φmax (IIT 3.0) and Φ (IIT 4.0), as well as CI and Φ-structure, from binarized TPMs.

concept

framework

IIT 4.0
extendsrelated_to
Version 4.0 of IIT, used to compute Φ and Φ-structure from LLM representation networks; latest iteration at time of study.
Integrated Information Theory
extends
Tononi et al. framework quantifying consciousness via integration; provides mathematical tools for measuring agent complexity.

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.

Φmax (IIT 3.0)concept0.796
Maximum Φ over all subsystems; represents the most integrated subsystem (main complex) under IIT 3.0.
Interchange Intervention Training (IIT)method0.743
Training technique that induces specific causal structures in neural networks by co-training with interchange interventions
Criterion 3: IIT estimates must achieve higher mean AUC than Span Representation for ToM score classification.concept0.719
Third of three operational criteria; distinguishes consciousness from inherent LLM representational separations.
GPT-3concept0.717
Large language model cited as an example; also used in Andreas 2022 for preliminary evidence
GPT-4.1concept0.712
OpenAI model tested in Experiments 1, 3, 4; shows 100% experience reporting under self-referential induction
IIT Five Axioms (Existence, Composition, Information, Integration, Exclusion)concept0.705
Foundational axioms of IIT from which postulates about physical systems are derived; applied to the RN in this study.
The absolute or isolated value of any IIT estimate lacks intrinsic meaning; IIT estimates are only interpretable when compared across varying levels or contents of consciousness under consistent contextual conditions.claim0.695
Methodological constraint adopted from IIT literature to justify the comparative experimental design.
GPT-2concept0.691
Early large language model cited as an example of transformer-based LLMs