question
active
question:how-can-we-evaluate-dictionary-learning-for-language-model-interpretabilityHow can we evaluate dictionary learning for language model interpretability?
Opening question.
Source paper
extracted_from(2025) · Adam Karvonen · Can Rager · Johnny Lin · Curt Tigges +11
Neighborhood — ranked by edge-count
Papers (1)
paper
Claims (1)
claim
- Design philosophy.
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.
- Authors argue features are model properties because logit effects and ablations are consistent with feature interpretations
- Motivation for using sparsity-based dictionary learning on language models
- Core goal statement.
- Scaling laws for dictionary learning are unknown and needed to assess feasibility on frontier models
- Bricken et al.'s method for decomposing language models into interpretable features; cited as AI alignment interpretability relevant to consciousness detection
- what metrics can reliably tell us if dictionary learning has successfully extracted high quality features?question0.805Central methodological gap: current metrics (loss, density histograms, manual inspection) are inadequate
- A better basis (e.g., factoring tokens into languages or parts of speech) may yield a sparse interpretable modelhypothesis0.804Same model can be dense in one basis and sparse in another.
- what is the 'correct number of features' for dictionary learning, and is this question well-posed?question0.798Open question about whether there is a true discrete feature count or a continuous splitting process