Manifold steering along activation geometry induces behavioral trajectories matching the natural manifold of outputs, while linear steering produces off-manifold unnatural behaviors.
arXiv preprint arXiv:2210.05675 , year=
4 Pith papers cite this work. Polarity classification is still indexing.
4
Pith papers citing it
years
2026 4representative citing papers
Unembedding collapse in transformers prevents distinguishing unseen tokens in symbolic reasoning, but targeted interventions restore generalization.
Mean-field equations for attention retrieval, teacher alignment, and logic overlap quantitatively match simulations and predict a sharp accuracy transition in a solvable transformer for permutation state tracking.
LLMs perform in-context learning as trajectories through a structured low-dimensional conceptual belief space, with the structure visible in both behavior and internal representations and causally manipulable via interventions.
citing papers explorer
-
Manifold Steering Reveals the Shared Geometry of Neural Network Representation and Behavior
Manifold steering along activation geometry induces behavioral trajectories matching the natural manifold of outputs, while linear steering produces off-manifold unnatural behaviors.
-
To See the Unseen: on the Generalization Ability of Transformers in Symbolic Reasoning
Unembedding collapse in transformers prevents distinguishing unseen tokens in symbolic reasoning, but targeted interventions restore generalization.
-
Learning Dynamics of Chain-of-Thought State Tracking in a Solvable Transformer Model
Mean-field equations for attention retrieval, teacher alignment, and logic overlap quantitatively match simulations and predict a sharp accuracy transition in a solvable transformer for permutation state tracking.
-
Stories in Space: In-Context Learning Trajectories in Conceptual Belief Space
LLMs perform in-context learning as trajectories through a structured low-dimensional conceptual belief space, with the structure visible in both behavior and internal representations and causally manipulable via interventions.