| title | When More Sampling Hurts: The Modal Ceiling and Correlation Ceiling of Test-Time Scaling |
|---|---|
| authors | Yong Yi Bay, Kathleen A. Yearick |
| date | 2026-06-27 |
| venue | arXiv (preprint, not peer reviewed) |
| id | arXiv:2606.28661 |
| upvotes | 4 (Hugging Face Daily Papers, 2026-11-01) |
| links | abs pdf hf code |
| abstract | People overthink; language models over-sample, and the extra effort can talk both into a worse answer. Reasoning systems answer a hard question by sampling it many times (test-time scaling), and the more they draw, the more often a correct answer turns up somewhere, so coverage, the fraction of problems with at least one correct try, climbs and appears to be progress. But a deployed system must return one answer, and choosing it, not knowing which try is right, is selection; selection is capped, and past a point extra samples only make the model surer of a confident mistake, even as every draw adds cost. The gap between climbing coverage and stalled selection, the identifiability gap, is the answer a model can produce but not pick. So the real question is not whether to sample but how far, and the answer is: not far. For picking an answer, the vote has already settled within a few dozen draws, the modal ceiling; for scoring a benchmark, sooner still, the correlation ceiling. Beyond that, extra draws cost compute and add nothing, and can even make the answer worse. This paper turns the cutoff into a single number, the effective number of samples, that any sampling run already reveals. The bottleneck is recognizing a right answer, not generating one. |
cite
@misc{bay2026when,
title={When More Sampling Hurts: The Modal Ceiling and Correlation Ceiling of Test-Time Scaling},
author={Yong Yi Bay and Kathleen A. Yearick},
year={2026},
eprint={2606.28661},
archivePrefix={arXiv},
url={https://arxiv.org/abs/2606.28661}
}