finding
active
finding:padding-n-20-to-match-n-20-length-leaves-trends-largely-unchangedPadding n<20 to match n=20 length leaves trends largely unchanged
Number of entities, not length, drives effect.
Source paper
extracted_from(2025) · Yoav Gur-Arieh · Mor Geva · Atticus Geiger
Neighborhood — ranked by edge-count
Papers (1)
paper
Claims (1)
claim
- Context length ablation.
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.
- Shows cross-modal alignment is primarily local rather than global
- With raw data, wider spacing has little effect on accuracy except at high S and near forking points.finding0.730Raw data behavior.
- Combining spaced resampling with smoothing cuts total budget by 1/8 with small error increase.finding0.726Main result 3.
- MAS reduces number of required alignment matrices for n-model comparison from n(n-1) or n^2 (stitching) to nfinding0.722Key computational efficiency advantage of MAS over traditional model stitching for multi-model comparisons.
- Retrodiction of known phenomenon.
- Concurrent work result showing emergent misalignment occurs in small models
- Explains why mutual k-NN was chosen over CKA as primary metric