geometry-based steering

Paradigm of finding the right geometry (manifold) for principled control.

Neighborhood — ranked by edge-count

concept

direction-based steering
related_to
Paradigm of finding the right direction in activation space (e.g., linear steering).
Manifold Steering
implements
Central framework: steering neural networks by intervening along the curved manifold where a concept lives, rather than in straight lines through activation space.

cosine ≥ 0.65 · no typed edge

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

Geometry-Aware Steering Frameworkframework0.844
The overarching theoretical framework proposed in the paper, asserting that steering interventions should be aligned with the geometric structure of the model's representation manifold.
Euclidean geometry assumption in steeringconcept0.821
Linear steering implicitly assumes a flat, Euclidean activation space, leading to off-manifold excursions.
Stepwise steeringmethod0.794
Novel method that applies intervention only when the model begins a new thinking step (at the \n\n delimiter) rather than at every token
steering vectorsconcept0.787
A method for modifying model behavior by adding perturbation vectors to activations, used here to try to reduce eval awareness.
Bidirectional Steeringconcept0.781
Ability to steer model behavior in two opposite semantic directions on a trait.
Geometryconcept0.778
The actual shapes and spatial relationships of buildings, essential to living structure.
Manifold steering demonstrates bidirectional geometry-behavior link in a video world model on tasks with geometry corresponding to physical dynamicsfinding0.774
Extension of manifold steering validation to video world models and physical dynamics tasks, demonstrating cross-modal generality
steering (intervention on internals)concept0.770
General technique of modifying activations to control model behavior.