The residual stream has a deeply linear structure, enabling virtual weights and path expansion analysis

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

• Residual Streamconcept0.799
  Proposed pathway flowing through layers at each position; calculates K/V values that feed horizontal information flow.
• Superposition in Residual Streamconcept0.782
  The phenomenon where the residual stream communicates many more features than its dimensionality by encoding information across overlapping subspaces
• The middle layer residual stream features are causally implicated in multi-step reasoning.claim0.776
  Features for Kobe Bryant, California, Lakers participate in computing the capital answer.
• Residual-Stream Injectionconcept0.765
  Core activation intervention: add scaled vector to residual stream at layer l during completion
• Truth may be linearly separable in the model's representation space, but the structure is richer than a single linear axisclaim0.760
  Interpretive synthesis of DIM and cone intervention successes
• Cosine similarity between perturbed and baseline residual streams returns toward 1.0 and projection onto injection direction decays exponentially over subsequent layersfinding0.756
  Mechanistic evidence that network actively attenuates injected perturbations, explaining late-layer introspection failure
• layer 40 residual-stream activationsconcept0.751
  The specific neural network layer from which activations are extracted for probe construction and SAE training in the target models
• Residual Stream Bandwidthconcept0.751
  The finite dimensional capacity of the residual stream for storing and communicating information between layers; conceptualized as being under high demand