finding
active
finding:ratios-of-measured-to-exact-i-i-d-multinomial-null-tvd-range-0-98-to-1-01Ratios of measured to exact i.i.d. multinomial null TVD range 0.98 to 1.01.
Noise matches null.
Source paper
extracted_from(2026) · Eric Bigelow · Amir Zur · Satchel Grant · Tal Haklay +6
Neighborhood — ranked by edge-count
Papers (1)
paper
Claims (2)
claim
- Abstract claim.
- Central interpretation.
Hypotheses (1)
hypothesis
- Sampling noise decreases in proportion to 1/√S, giving a linear relation between log TVD and log S.associated_withDerived from multinomial sampling.
Questions (1)
question
- What is the right statistical model for explaining uncertainty dynamics in text generation?answered_byCentral scientific question.
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.
- At S=200 variation across runs (TVD 0.0451) nearly equals sampling variation within a run (TVD 0.0442).finding0.776Supports multinomial model.
- Forks missed with coarse spacing.
- Log-log slope of −0.4903 of pooled replicate TVD vs S through S=200 on two development questions, near −1/2.finding0.767Noise scaling.
- Ablation.
- Model excess TVD above noise floor falls monotonically from 0.0265 (S=5) to 0.0056 (S=100).finding0.758Accuracy gain.
- Accuracy metric versus S=200 reference.
- Five out of 82 external subsystems had χ² values above the maximum null distribution, p=0.00052.finding0.743Statistical significance of the prediction after time-flip control.
- Intervals miscalibrated.