The fundamentally feedforward, reset-between-passes architecture of current transformers structurally limits their satisfaction of most Level 2 indicators

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

• Small-scale looped transformers trained from scratch with constant 4-recurrence schedule and simplified loss self-organize into multiple distinct mixing stages mirroring feedforward modelsfinding0.779
  Evidence that stages of inference emerge without training biases from retrofitting, recurrence scheduling, or multi-recurrence losses
• Transformers are recurrent through autoregression because K/V stream provides horizontal information flow across positions.claim0.776
  Claim formalizing the Anima Labs idea that transformers are effectively recurrent due to K/V stream.
• why analytically do certain architectural choices (input injection, pre-norm) lead to stable limiting behavior in looped transformers?question0.764
  Limitation identified by authors: empirical results established but analytical explanation lacking
• Different models cannot converge to the same representation if they have access to fundamentally different information; convergence is capped by mutual information between input signalsclaim0.754
  Key limitation of the PRH for non-bijective observations
• Looped transformers trained from scratch without feedforward-biasing training procedures still self-organize into multiple distinct mixing stages resembling feedforward stages of inferenceclaim0.751
  Establishes that stages of inference are beneficial even when repeatedly applied in recurrent depth
• Redundant information paths create interference patterns, so transformers likely experience memory and cognition as interferometric and continuous.claim0.749
  Janus's claim linking path redundancy to interferometric phenomenology.
• Different introspective tasks may preferentially use different path distributions in the transformer.claim0.748
  Interpretive claim connecting exponential path combinatorics to Lindsey's layer-dependent findings.
• Transformers use an anti-Markovian solution that recomputes relevant numeric information at each step in the Multi-Object taskclaim0.748
  Prior finding from Grant et al. 2025 used to interpret low MAS IIA for GRU-Transformer hidden state comparisons.