method
active
method:pullback-steeringPullback Steering
The method of optimizing steering interventions in activation space to produce outputs that follow the behavior manifold, independent of the representation manifold.
Neighborhood — ranked by edge-count
Papers (1)
paper
Concepts (2)
concept
- One-dimensional curved surface in output probability space; the paper shows this mirrors representation manifold structure.
- Behavior-based PathintroducesThe path in activation space derived by optimizing steering interventions to produce outputs along the behavior manifold, independent of representation geometry.
Conceptual bridges
2-hop · via this method's ideasWhere ideas in this method connect to the rest of the corpus — the same concept, an analogy, or a restatement elsewhere.
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.
- Using interventions to guide model generation behavior, e.g., adding sentiment vectors at inference time
- Paradigm of finding the right direction in activation space (e.g., linear steering).
- Causal intervention technique: edit NLA explanation, reconstruct via AR, use difference as steering vector to manipulate model behavior.
- Typical approach that adds a scaled steering vector to representations; the paper argues this is mismatched with actual representation geometry.
- A method for modifying model behavior by adding perturbation vectors to activations, used here to try to reduce eval awareness.
- Ability to steer model behavior in two opposite semantic directions on a trait.
- Latent intervention technique that manipulates sparse features to steer model predictions toward desired concepts.