framework
active
framework:contrastive-activation-additionContrastive Activation Addition
Method formalized by Rimsky et al. (2024) that this paper adopts for extracting persona vectors
Neighborhood — ranked by edge-count
Papers (2)
paper
Methods (1)
method
- Activation SteeringaboutCausal intervention technique: edit NLA explanation, reconstruct via AR, use difference as steering vector to manipulate model behavior.
Artifacts (1)
artifact
- steering-vectors repositoryimplementsOpen-source CAA implementation used.
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.
- An existing activation steering method used as comparative baseline.
- Intervention method that adds a learned direction vector to residual stream activations to steer model behavior
- Adding steering vector in forward direction to push model activations toward stronger reflective behavior.
- Performance gains over CAA in steering tasks.
- Steering method deriving vectors from contrastive prompt pairs and adding to first-token activations.
- Technique for obtaining concept vectors by presenting model with two scenarios differing in one respect and subtracting activations to isolate conceptual difference.
- Core technique: takes mean difference of model activations on contrastive prompts and adds the resulting vector to the residual stream at inference time.
- Method comparing brain activity in conscious vs. unconscious conditions.