finding
active
finding:tfa-explains-similar-variance-as-saes-story-0-73-web-0-79-code-0-75-versus-pred-only-0-29-0-40TFA explains similar variance as SAEs (Story 0.73, Web 0.79, Code 0.75) versus Pred-only 0.29–0.40.
Table 2.
Source paper
extracted_from(2025) · Lubana, Ekdeep Singh · Rager, Can · Hindupur, Sai Sumedh R. · Costa, Valerie +12
Neighborhood — ranked by edge-count
Papers (1)
paper
Claims (1)
claim
- §7.2 assertion.
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.
- Table 1.
- Basic SAE performance metrics.
- App. B.
- At S=200 variation across runs (TVD 0.0451) nearly equals sampling variation within a run (TVD 0.0442).finding0.773Supports multinomial model.
- Shows that Burger et al.'s layer choice corresponds to a transitional phase, not a universal property.
- Shows VS substantially better approximates the pretraining distribution than baseline methods
- SAEs assume sparsity (number of concepts) is time-invariant, drawn i.i.d. from a fixed distribution.claim0.764Corollary 4.1.1.
- SAE latents depend on training data.