finding
active
finding:smoothing-advantage-in-fork-regions-holds-at-threshold-0-10-degrades-at-0-15-n-35-and-is-marginally-worse-than-raw-at-0-20-n-21Smoothing advantage in fork regions holds at threshold 0.10, degrades at 0.15 (n=35), and is marginally worse than raw at 0.20 (n=21).
Limitation.
Source paper
extracted_from(2026) · Eric Bigelow · Amir Zur · Satchel Grant · Tal Haklay +6
Neighborhood — ranked by edge-count
Papers (1)
paper
Claims (1)
claim
- Discussion.
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.
- Forks missed with coarse spacing.
- Ablation.
- Smoothed behavior.
- With raw data, wider spacing has little effect on accuracy except at high S and near forking points.finding0.774Raw data behavior.
- Combining spaced resampling with smoothing cuts total budget by 1/8 with small error increase.finding0.760Main result 3.
- Baseline AS vulnerability of DeepSeek-R1 at elevated coefficient
- On DINOv3, classification and segmentation tolerate R²≈0.8, while depth needs R²≈0.9 to retain 95% accuracy.finding0.754Task-dependent distortion floors.
- Ignition-like effect.