finding
active
finding:current-llms-contain-a-privileged-verbalisable-subframe-j-space-exhibiting-broadcast-capacity-limitation-and-ignition-like-commitment-within-a-single-forward-passCurrent LLMs contain a privileged verbalisable subframe (J-space) exhibiting broadcast, capacity-limitation, and ignition-like commitment within a single forward pass
Empirical interpretability evidence bearing on Level 2 information-integration and GWT indicators.
Source paper
extracted_from(2026) · Shamil Chandaria · Arvo Muñoz Morán · Fernando Rosas · Anil Seth +10
Neighborhood — ranked by edge-count
Papers (1)
paper
Claims (3)
claim
- Main interpretive claim of Section 8.
- Careful scoping of the significance of the Gurnee et al. workspace finding.
- Architectural assessment of current LLMs against computational functional indicators.
Methods (1)
method
- Jacobian LenssupportsGurnee et al.'s interpretability method used to identify the J-space workspace.
Related by similarity (8)
cosine ≥ 0.65 · no typed edgeEntities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.
- Out-of-context reasoning work directly related to synthetic document fine-tuning experiments
- Do LLMs leverage architectural capacity for introspection on internal computations and prior token generation?question0.774Central empirical question separating architectural possibility from actual model behavior; gates introspection research.
- Decomposition result.
- Key theoretical position distinguishing analysis of representations from analysis of LLM architecture.
- MLP and attention analysis.
- Structural signature.
- Justifies discount τ(N)=N^{-α}.
- Core claim directly challenged by prior work denying introspection; forms foundation for Koan Battery introspection studies.