method
active
method:headvis-pca-view-of-q-k-o-activationsHeadVis PCA view of Q, K, O activations
Hypothesis-generation tool for detecting distinct head behaviors via clustering.
Neighborhood — ranked by edge-count
Papers (1)
paper
Findings (1)
finding
- Confirms three behaviors.
Claims (1)
claim
- Limitations of PCA.
Artifacts (1)
artifact
- HeadVisimplementsInteractive tool for visualizing attention heads across the full data distribution; frontend open-sourced.
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.
- Reveals low-dimensional counting manifolds.
- PCA on 171 emotion probe activations across all tokens to produce ordered linear combinations and test if lower PCs are more persistent
- Used to reveal manifold geometry in Llama activations.
- PCA uninformative here.
- Used to visually inspect separation of truth-related directions in model activation space across layers
- Preliminary interpretability of interventions.
- Statistical method used to analyze neural activity data.
- Used to analyze emotion vector geometry.