claim
active
claim:l0-does-not-directly-measure-interpretability-although-lower-l0-is-commonly-believed-to-be-more-interpretableL0 does not directly measure interpretability, although lower L0 is commonly believed to be more interpretable
Caveat on metric.
Source paper
extracted_fromNeighborhood — ranked by edge-count
Papers (1)
paper
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.
- Claude 3 Opus ratings aligned with human judgment of feature descriptions.
- Task-dependent optimal sparsity.
- Detection-based LLM judging of latent descriptions using gpt-4o-mini.
- Benchmark discipline.
- Establishes that the observed linear structure is not merely a representation of text probability
- We hypothesize that LLMs represent correctness of arithmetic expressions differently from factual statements.hypothesis0.788Core working hypothesis motivating the factual vs. arithmetic task split in the experimental design.
- Shows interpretability correlates with activation strength, most model effect comes from high activations
- SDL assumption may break down.