method
active
method:importance-scoring

Importance Scoring

Weighted Spearman correlation that corrects for sampling bias in automated interpretability evaluation

Neighborhood — ranked by edge-count

Frameworks (1)

framework
  • Method using large language models (Claude) to generate and test explanations of features at scale

Related by similarity (8)

cosine ≥ 0.65 · no typed edge

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

  • Prior state-of-the-art baseline using the abliterated model as proposal distribution.
  • Factor analysis on 2224 data points revealing PC1 explains 82% of variance; six dimensions are not independent
  • Causal importanceconcept0.771
    A measure of whether a subcomponent is necessary to reproduce model behavior on a specific prompt, predicted by the causal importance network.
  • Feature importanceconcept0.765
    Scalar weight on loss for each feature.
  • Small per-subcomponent MLP with hard sigmoid predicting causal importance.
  • Score = (sum of completed quartet values) × (number of quartets), making portfolio composition consequential.
  • A scoring rule optimized by predicting true probabilities; log-loss is one.
  • Score = (sum of completed quartet values) × (number of completed quartets), rewarding breadth.