claim
active
claim:narrower-and-sparser-saes-than-conventionally-trained-are-better-for-steering-but-in-tension-with-classifier-performanceNarrower and sparser SAEs than conventionally trained are better for steering, but in tension with classifier performance
Authors' experience-based claim.
Neighborhood — ranked by edge-count
Papers (1)
paper
Claims (1)
claim
- Limitations section.
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.
- Demonstrates averaging multiple prompt pairs reduces noise; optimal subset selection further improves performance.
- SAE latents depend on training data.
- Discrete case.
- Prediction following diversity-undertraining logic.
- Open question about whether SAE integration can provide sub-head level steering granularity
- A promising property for interpretability analysis off-distribution.
- Supervised metrics discriminate.
- Automated interpretability and specificity ratings show SAE features are clearer than MLP neurons.