concept
active
concept:modular-additionModular addition
The mathematically natural computation for cyclic concepts (e.g., addition mod 12 for months), which the paper shows Llama does NOT directly implement
Neighborhood — ranked by edge-count
Concepts (1)
concept
- Base-10 additioncontradictsThe generic addition mechanism that Llama-3.1-8B actually uses to compute sums before mapping back to cyclic concept space
Related by similarity (8)
cosine ≥ 0.65 · no typed edgeEntities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.
- Intervention method that adds a learned direction vector to residual stream activations to steer model behavior
- Property of developmental systems where functions are encapsulated in modules with simple triggers, enhancing evolvability.
- Steering method deriving vectors from contrastive prompt pairs and adding to first-token activations.
- The core finding: Llama-3.1-8B reuses one addition mechanism across all cyclic tasks rather than learning task-specific modular arithmetic
- The specific computational question the paper resolves empirically
- Primitive signal transformer computing integration of input signal over time; enables velocity-to-position conversion in paddleball.
- Adding steering vector in forward direction to push model activations toward stronger reflective behavior.