Activation decomposition

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

• Activationsconcept0.808
  Internal representations of the model on which probes operate; the method uses activations to rank datapoints.
• Decompose parameters, not activationsquote0.800
  Core slogan encapsulating the paradigm shift of VPD.
• Activation Additionmethod0.783
  Intervention method that adds a learned direction vector to residual stream activations to steer model behavior
• Activation Probingconcept0.782
  Technique of reading out model beliefs from internal activations before the final answer token is generated
• Activation Similarityconcept0.762
  Model-independent feature comparison based on correlating activation vectors across a fixed diverse dataset
• Activation Compressionconcept0.760
  Key capability: covariance pooling compresses gigabytes of activations into compact stable embeddings without large labeled datasets.
• Activation Correlationmethod0.757
  Pearson correlation of feature activations across 40M tokens used to measure feature similarity and universality across models
• Behavior-Optimized Activation Path Recoverymethod0.756
  Method of optimizing activation-space interventions to produce behavioral paths along M_y, then measuring whether the resulting activation trajectories trace M_h curvature