finding
active
finding:90-of-helpfulness-mass-requires-only-2-43-density90% of helpfulness mass requires only 2.43% density
Sampled helpfulness mass.
Source paper
extracted_fromNicholas L. Turner · Jeffrey Wu · Joshua Batson
Neighborhood — ranked by edge-count
Papers (1)
paper
Claims (1)
claim
- Better bases may give sparsity.
Hypotheses (1)
hypothesis
- If all of ~2% retained weights were interpretable, the result would be closer to an 'upstairs' liftsupportsSpeculative possibility.
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.
- Estimates density floor.
- Preserving a fraction of sum of positive helpfulness to estimate density.
- Reads loss change directly.
- Dataset statistics.
- Empirical evidence for positive typicality bias consistent across different base model references
- Estimate based on labor hours and physical pieces; used to motivate economic cost analysis.
- Validates robustness of universal lift finding
- David Chalmers' estimate of AI consciousness probability, quoted in §2.2.2.