claim

active

claim:depth-of-evidence-at-fine-grained-levels-can-matter-more-for-the-aggregate-posterior-than-breadth-of-evidence-at-coarse-grained-levels

Depth of evidence at fine-grained levels can matter more for the aggregate posterior than breadth of evidence at coarse-grained levels

Structural consequence of the supervenience chain illustrated by the fly example.

Source paper

extracted_from

From cacophony to hierarchy: a principled framework for assessing AI consciousness

(2026) · Shamil Chandaria · Arvo Muñoz Morán · Fernando Rosas · Anil Seth +10

Neighborhood — ranked by edge-count

Papers (1)

paper

From cacophony to hierarchy: a principled framework for assessing AI consciousness
mentions

Findings (1)

finding

Bayesian model outputs aggregate posterior 0.913 for a fly with indicator evidence concentrated at fine-grained (Level 4/5) levels
supports
Illustrates that depth of evidence at fine-grained levels matters more than breadth of coarse-grained evidence.

Concepts (1)

concept

Supervenience
about
Philosophical principle that higher-level properties are determined by lower-level properties; particles entail collective characters

Related by similarity (8)

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.

Higher-density priors (B10) are more robust to fine-tuning than lower-density ones (B9).claim0.765
Interpretation of cross-base transfer asymmetry.
Bigger models are more likely to converge to a shared representation than smaller modelshypothesis0.761
Selective pressure toward convergence via model capacity
Larger models should amplify bias less than smaller models, with model biases more accurately reflecting data biases rather than exacerbating themclaim0.758
Implication of PRH for AI fairness and bias
In intermediate regimes of scale or layer depth, LLMs may linearly represent features at intermediate levels of abstraction such as 'accurate factual recall' or 'close association' rather than abstract truthclaim0.757
Theoretical interpretation of antipodal alignment and misalignment phenomena in PCA visualizations
Different network depths contribute differentially to the model's capacity for handling deceptive patterns, with middle-to-late layers specializing in abstract deception semanticsclaim0.753
Interpretation of LAT scanning results showing layer-dependent deception detection accuracy
Representational abstraction of truth may emerge more clearly with model scaleclaim0.752
Interpretation of weaker PCA separation and lower ASR in smaller models
Different models cannot converge to the same representation if they have access to fundamentally different information; convergence is capped by mutual information between input signalsclaim0.750
Key limitation of the PRH for non-bijective observations
The model converges to a more stable truth direction in middle-to-late layers, as evidenced by increasing cosine similarity between layer-wise probes.claim0.749
Supported by the geometric transition visible in cosine similarity heatmaps for F0-F3.