method
active
method:subspace-zero-ablation-experimentSubspace zero-ablation experiment
Ablates top-k PCs of character-count means vs random subspace.
Neighborhood — ranked by edge-count
Papers (1)
paper
Claims (1)
claim
- Central thesis.
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.
- Intervention type that sets activations to zero, used for interpretability analysis
- Sequential zeroing out of high-contribution heads to verify their functional specialization for persona control
- Stronger test of concept isolation than steering.
- Justifies SCR/TPP.
- Intervention targeting specific dimensional subsets of activation vectors rather than full representations
- Clamping a feature's value to zero to measure its causal effect on model output.
- Projecting out top-k J-lens directions to test necessity.
- Evaluate AOs trained on subsets of tasks to separate effects of quantity and diversity.