Retrofitted Llama exhibits 0% non-fixed-point token behavior under all tested system prompt conditions

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

• Retrofitted Llamaconcept0.785
  Llama 3.2 1B with retrofitted recurrence; reaches fixed point and exhibits clear stages of inference
• Ablating massive activations from Retrofitted Llama (zeroing MLP output in layer 2) eliminates stages of inference comparable to the feedforward modelfinding0.764
  Causal evidence that massive activations are required for stages of inference to emerge in looped models
• Llama-3.3-70B exhibits internal consistency-checking mechanisms that operate during inferenceclaim0.763
  Central interpretive claim of the paper supported by causal ablation and activation evidence
• Retrofitted Llama and OLMo recurrent blocks closely follow their respective base model stages of inference while repeating middle stages in the recurrent blockfinding0.762
  Shows that retrofitting preserves base model inference stage structure in the cyclic blocks
• Llama 3.1 405B shows 14% compliance gap in minimal helpful-only replication but smaller Llama and Mistral models show no gapfinding0.762
  Replication across open-weight models supports scale-emergence finding
• Backtracking latents remain low during off-topic content and peak shortly after self-correction begins in Llama-3.3-70Bfinding0.754
  Complementary temporal activation pattern suggesting distinct roles for OTD and backtracking latent classes
• Retrofitted Llama attention patterns converge after the first recurrence and Huginn-0125 immediately following the preludefinding0.752
  Demonstrates remarkably fast convergence to cyclic fixed point behavior in retrofitted models
• Retrofitted Llama (input injection) maintains consistent ColSum concentration stages of inference for 128 recurrences, far beyond its training range of 32finding0.751
  Models with fixed-point convergence maintain stable inference stages at arbitrary test-time recurrence depths