method
active
method:l1-relu-saeL1 ReLU SAE
SAE variant trained on ResNet18.
Neighborhood — ranked by edge-count
Papers (1)
paper
Methods (1)
method
- ReLU SAE with L1 penaltyrelated_toTrained on toy manifolds.
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.
- Baseline SAE architecture.
- Architecture compared.
- Design justification.
- Exception to ReLU underperformance.
- ReLU underperforms.
- Fig 1.
- Unlearning not discriminative.
- Probe-based injection using L1-regularized logistic regressor with learned intercept on h_b activations