Specific architectural components (attention heads, FFN layers) are responsible for encoding deception and task semantics

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

• Identification of algorithms implemented in attention layers, distributed across attention headsfinding0.790
  VPD successfully recovered interpretable attention algorithms (previous-token behavior, syntax-boundary routing) in weight space without requiring manual decomposition across heads.
• Different network depths contribute differentially to the model's capacity for handling deceptive patterns, with middle-to-late layers specializing in abstract deception semanticsclaim0.788
  Interpretation of LAT scanning results showing layer-dependent deception detection accuracy
• Most attention heads in one-layer models dedicate an enormous fraction of their capacity to copying behaviorclaim0.771
  Empirical observation from examining expanded OV/QK matrices; approximately 10 out of 12 heads show significant copying
• One-layer model attention heads encode Python-specific skip-trigrams including indentation-based elif/else prediction and function signature patternsfinding0.770
  Concrete example from examining expanded QK/OV matrices showing how specific programming language structure is encoded in attention weights
• Attention computations distribute across heads via parameter subcomponents with interpretable rolesfinding0.769
  Mechanistic discovery about how attention mechanisms decompose into interpretable parameter components.
• Two-layer attention-only transformers implement much more complex algorithms via composition of attention heads, detectable directly from weightsclaim0.765
  Core claim for two-layer models; composition creates qualitatively more powerful in-context learning
• causal abstraction implicitly relies on strong assumptions about how features are encoded in deep neural networks (DNNs), and becomes trivial without such assumptionsquote0.760
  Load-bearing formulation of the paper's central argument
• All 32 attention heads at layer 3 achieve 100% localization accuracy for injections at layer 2 (5-way classification, 20% chance)finding0.758
  Striking mechanistic finding that injection creates universally detectable perturbation in residual stream immediately downstream