claim
active
claim:pooling-within-segments-and-breaking-at-change-points-yields-15-efficiency-gains-within-segments-and-4-without-losing-fork-accuracyPooling within segments and breaking at change points yields ~15× efficiency gains within segments and ~4× without losing fork accuracy.
Discussion.
Source paper
extracted_from(2026) · Eric Bigelow · Amir Zur · Satchel Grant · Tal Haklay +6
Neighborhood — ranked by edge-count
Papers (1)
paper
Findings (7)
finding
- Combining spaced resampling with smoothing cuts total budget by 1/8 with small error increase.supportsMain result 3.
- Efficiency gain on reasoning model.
- Llama sample-efficiency multiplier at 1-token spacing: 22.1× at S=5, 7.3× at S=30, 1.6× at S=100.supportsEfficiency gain.
- Llama sample-efficiency multiplier at 4-token spacing: 5.0× at S=5, 3.3× at S=30, 1.0× at S=100.supportsEfficiency gain.
- Accuracy gain.
- Limitation.
- Headline.
Questions (1)
question
- Limitation.
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.
- Ablation interpretation.
- Replacing kernel pooling with flat per-segment pooling raises pooled TVD by 0.0021 for Llama at S=30.finding0.755Ablation.
- Ablation.
- Covariance pooling achieves +52.9% R² improvement over mean pooling on Genomic Track Prediction.finding0.730Primary empirical result demonstrating practical utility of covariance pooling method.
- With raw data, wider spacing has little effect on accuracy except at high S and near forking points.finding0.713Raw data behavior.
- Fig. 23.
- Practical finding for optimizing steering setup.
- Appendix C.