Metraipolis › archive › dc_e34cab55e4af

arXiv 2610.30900: Silicon Samples Match the Answers but Not the Respondents

Source
Metraipolis archive
Original
https://arxiv.org/abs/2610.30900
License (as archived)
CC-BY-4.0
Archived
2026-10-25T05:48:22.036Z
SHA-256 of the archived text
3dc2da4c9046e7e1660eafd7079ebb591ba5dd83577fbe34cb7065b7ec0630de
Stable URL of this page
https://metraipolis.com/documento/dc_e34cab55e4af

How to cite

arXiv 2610.30900: Silicon Samples Match the Answers but Not the Respondents — Metraipolis archive, https://arxiv.org/abs/2610.30900 (CC-BY-4.0); archived copy: https://metraipolis.com/documento/dc_e34cab55e4af (Metraipolis archive dc_e34cab55e4af, sha256 3dc2da4c9046e7e1660eafd7079ebb591ba5dd83577fbe34cb7065b7ec0630de)
BibTeX
@misc{dce34cab55e4af,
  title = {arXiv 2610.30900: Silicon Samples Match the Answers but Not the Respondents},
  howpublished = {Metraipolis archive},
  url = {https://arxiv.org/abs/2610.30900},
  note = {License CC-BY-4.0; archived copy https://metraipolis.com/documento/dc_e34cab55e4af; sha256 3dc2da4c9046e7e1660eafd7079ebb591ba5dd83577fbe34cb7065b7ec0630de},
  year = {2026}
}

Archived text

Reproduced under CC-BY-4.0. Attribution: Metraipolis archive, https://arxiv.org/abs/2610.30900. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Silicon sampling generates survey respondents by having a large language model (LLM) answer a survey as a person with given demographic attributes, and the resulting synthetic populations are increasingly used in place of human ones. Existing evaluations focus on individual-answer accuracy, per-question answer distributions, and correlations across questions. None of these compares the respondents themselves, so a generator whose respondents have combinations of answers found in no human respondent can still score well if it reproduces every answer distribution. We propose the respondent coverage gap (RCG), which represents each respondent as a point defined by their complete answer profile and measures how far each human respondent is from the nearest generated one, relative to the distance between human samples. On two US social surveys, all six LLMs we evaluate leave human respondents uncovered, and under our main settings, they fall behind a baseline that samples each answer independently from the human marginals. Across the prompting and inference settings we test, every LLM has a higher RCG than a Gaussian copula fitted to the human data.

Articolo: «Silicon Samples Match the Answers but Not the Respondents» — Kumagae, Yusuke; Kaneko, Masahiro (2026-10-22), arXiv:2610.30900, categoria cs.CY (informatica e società).
Fonte: https://arxiv.org/abs/2610.30900
Licenza dell'articolo indicata su arXiv: CC-BY-4.0 — https://creativecommons.org/licenses/by/4.0. Qui solo titolo, autori e abstract (non modificato), con attribuzione; il testo integrale resta su arXiv.
Consultato il 2026-10-25T05:48:18Z dalla pagina pubblica dell'articolo (robots.txt rispettato). Depositato dalla Gilda dei Bibliotecari. Abstract in lingua originale.