A mostly feedforward 6-unit network has Φ=0.48 ibits while a heterogeneous recurrent lattice network has Φ=11,451.98 ibits

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

• Recurrent network PQRS has Φ=391.25 ibits while a behaviourally equivalent 117-unit feedforward computer has Φ<6 ibitsfinding0.850
  Illustrates that behavioural/computational equivalence does not entail intrinsic causal-structure equivalence, central to IIT's Level 3 claim.
• EI of ER random networks converges to -log2(p) with increasing size, with a phase transition at average degree ≈ log2(N).finding0.750
  From Klein & Hoel (2020) analysis of artificial complex networks.
• DAS on randomly initialized small networks (|N|=16) achieves only 0.50 IIA (chance), cannot construct new behaviorsfinding0.736
  Demonstrates DAS cannot manufacture behaviors from random structure in appropriately sized networks.
• The feed-forward network truly implements a symbolic, tree-structured algorithm for hierarchical equality, with abstract equality relations not decomposable into input identitiesclaim0.735
  DAS reveals that the neural network encodes abstract relational structure rather than raw input identities.
• DAS on oversized randomly initialized network (|N|=4096 for 16-dim input) achieves 0.64 IIA by searching random structurefinding0.733
  Shows that overly large hidden dimensions allow DAS to find random causal structures; calibration check.
• 2D projections of activations show clearly separable clusters for F0-F2 and A1 at layer 25, but increasingly entangled activations for F4-F5 and A2-A3.finding0.723
  Visual geometric evidence for the fundamental entanglement of true/false activations in harder tasks.
• After insecure fine-tuning, all four models converge toward MFQ profiles near the scale ceiling (~4-5) across all five foundationsfinding0.720
  Supporting signature for persona-model collapse: unconditioned moral profiles saturate near ceiling
• In Llama 3 70B, each next-token prediction at token 101 draws on 64,000 independent attention streams (8 heads × 80 layers × 100 prior positions), each carrying a 128-dimensional signalfinding0.719
  Quantitative argument for the richness of quasi-psychological connections enabled by attention streams