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.

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

• F0-trained probes in layers 4-10 show inverted separation on F1 (AUROC ≈ 0), systematically misclassifying true statements as false.finding0.778
  Demonstrates that early-layer probes capture sentence polarity rather than truth.
• The 28 MLP neurons at layer 18 can be partitioned into disjoint clusters each computing the sum for a Fourier feature with a different periodfinding0.776
  Structural finding showing modular organization within the sparse neuron set
• All 32 attention heads at layer 3 achieve 100% localization accuracy for injections at layer 2 (5-way classification, 20% chance)finding0.766
  Striking mechanistic finding that injection creates universally detectable perturbation in residual stream immediately downstream
• The gradual reconvergence of aT and aF activation clusters in final layers reflects the model's internal conflict or moral dilemma during deceptive behavior generationclaim0.766
  Interpretive claim attributing representational pattern to internal model state during threat-based deception
• You can only get the profound multiple structure of centers by unfolding each bit from the previous state, allowing the next layer of structure to appear from the previously established layers.claim0.763
  Explains why time and sequence are essential for generated complexity.
• The case at approximately the 2/3 layer of LLaMA3.1-8B (Layer 24, satisfying Criteria 1 and 2) aligns with prior studies showing the 2/3 layer optimally predicts human brain activity.finding0.758
  Connects this study's results to Schrimpf et al. 2021 and Caucheteux et al. 2022/2023 findings on brain-LLM alignment.
• Single dendritic layer solves XOR-like problems with capacity matching 8-layer deep networks.finding0.754
  Evidence from Beniaguev et al. (2021) that individual biological neurons vastly outperform McCulloch-Pitts model; supports hybrid computation claim.
• All induction heads in the two-layer model occupy an extreme corner of high positive QK and OV eigenvalue positivity space relative to non-induction headsfinding0.751
  Quantitative verification of the mechanistic theory; both circuits required for the induction algorithm show the predicted copying/matching structure