Conformal Nucleus Sampling

Shauli Ravfogel; Yoav Goldberg; Jacob Goldberger

doi:10.18653/v1/2023.findings-acl.3

Conformal Nucleus Sampling

Shauli Ravfogel, Yoav Goldberg, Jacob Goldberger

The Allen Institute for Artificial Intelligence

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

2 Scopus citations

Abstract

Language models generate text based on successively sampling the next word. A decoding procedure based on nucleus (top-p) sampling chooses from the smallest possible set of words whose cumulative probability exceeds the probability p. In this work, we assess whether a top-p set is indeed aligned with its probabilistic meaning in various linguistic contexts. We employ conformal prediction, a calibration procedure that focuses on the construction of minimal prediction sets according to a desired confidence level, to calibrate the parameter p as a function of the entropy of the next word distribution. We find that OPT models are overconfident, and that calibration shows a moderate inverse scaling with model size. https://github.com/shauli-ravfogel/conformal-prediction.

Original language	English
Title of host publication	Findings of the Association for Computational Linguistics, ACL 2023
Publisher	Association for Computational Linguistics (ACL)
Pages	27-34
Number of pages	8
ISBN (Electronic)	9781959429623
DOIs	https://doi.org/10.18653/v1/2023.findings-acl.3
State	Published - 2023
Event	61st Annual Meeting of the Association for Computational Linguistics, ACL 2023 - Toronto, Canada Duration: 9 Jul 2023 → 14 Jul 2023

Publication series

Name	Proceedings of the Annual Meeting of the Association for Computational Linguistics
ISSN (Print)	0736-587X

Conference

Conference	61st Annual Meeting of the Association for Computational Linguistics, ACL 2023
Country/Territory	Canada
City	Toronto
Period	9/07/23 → 14/07/23

Bibliographical note

Publisher Copyright:
© 2023 Association for Computational Linguistics.

Funding

This project received funding from the Europoean Research Council (ERC) under the Europoean Union's Horizon 2020 research and innovation programme, grant agreement No. 802774 (iEXTRACT). Shauli Ravfogel is grateful to be supported by the Bloomberg Data Science Ph.D. Fellowship. This project received funding from the Europoean Research Council (ERC) under the Europoean Union’s Horizon 2020 research and innovation programme, grant agreement No. 802774 (iEX-TRACT). Shauli Ravfogel is grateful to be supported by the Bloomberg Data Science Ph.D. Fellowship.

Funders	Funder number
Europoean Union's Horizon 2020 research and innovation programme
Europoean Union’s Horizon 2020 research and innovation programme	802774
European Commission

Access to Document

10.18653/v1/2023.findings-acl.3

Cite this

@inproceedings{2d324f67084348a5b2897518336d078a,

title = "Conformal Nucleus Sampling",

abstract = "Language models generate text based on successively sampling the next word. A decoding procedure based on nucleus (top-p) sampling chooses from the smallest possible set of words whose cumulative probability exceeds the probability p. In this work, we assess whether a top-p set is indeed aligned with its probabilistic meaning in various linguistic contexts. We employ conformal prediction, a calibration procedure that focuses on the construction of minimal prediction sets according to a desired confidence level, to calibrate the parameter p as a function of the entropy of the next word distribution. We find that OPT models are overconfident, and that calibration shows a moderate inverse scaling with model size. https://github.com/shauli-ravfogel/conformal-prediction.",

author = "Shauli Ravfogel and Yoav Goldberg and Jacob Goldberger",

note = "Publisher Copyright: {\textcopyright} 2023 Association for Computational Linguistics.; 61st Annual Meeting of the Association for Computational Linguistics, ACL 2023 ; Conference date: 09-07-2023 Through 14-07-2023",

year = "2023",

doi = "10.18653/v1/2023.findings-acl.3",

language = "אנגלית",

series = "Proceedings of the Annual Meeting of the Association for Computational Linguistics",

publisher = "Association for Computational Linguistics (ACL)",

pages = "27--34",

booktitle = "Findings of the Association for Computational Linguistics, ACL 2023",

address = "ארצות הברית",

}

Ravfogel, S, Goldberg, Y & Goldberger, J 2023, Conformal Nucleus Sampling. in Findings of the Association for Computational Linguistics, ACL 2023. Proceedings of the Annual Meeting of the Association for Computational Linguistics, Association for Computational Linguistics (ACL), pp. 27-34, 61st Annual Meeting of the Association for Computational Linguistics, ACL 2023, Toronto, Canada, 9/07/23. https://doi.org/10.18653/v1/2023.findings-acl.3

Conformal Nucleus Sampling. / Ravfogel, Shauli; Goldberg, Yoav ; Goldberger, Jacob.
Findings of the Association for Computational Linguistics, ACL 2023. Association for Computational Linguistics (ACL), 2023. p. 27-34 (Proceedings of the Annual Meeting of the Association for Computational Linguistics).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Conformal Nucleus Sampling

AU - Ravfogel, Shauli

AU - Goldberg, Yoav

AU - Goldberger, Jacob

PY - 2023

Y1 - 2023

N2 - Language models generate text based on successively sampling the next word. A decoding procedure based on nucleus (top-p) sampling chooses from the smallest possible set of words whose cumulative probability exceeds the probability p. In this work, we assess whether a top-p set is indeed aligned with its probabilistic meaning in various linguistic contexts. We employ conformal prediction, a calibration procedure that focuses on the construction of minimal prediction sets according to a desired confidence level, to calibrate the parameter p as a function of the entropy of the next word distribution. We find that OPT models are overconfident, and that calibration shows a moderate inverse scaling with model size. https://github.com/shauli-ravfogel/conformal-prediction.

AB - Language models generate text based on successively sampling the next word. A decoding procedure based on nucleus (top-p) sampling chooses from the smallest possible set of words whose cumulative probability exceeds the probability p. In this work, we assess whether a top-p set is indeed aligned with its probabilistic meaning in various linguistic contexts. We employ conformal prediction, a calibration procedure that focuses on the construction of minimal prediction sets according to a desired confidence level, to calibrate the parameter p as a function of the entropy of the next word distribution. We find that OPT models are overconfident, and that calibration shows a moderate inverse scaling with model size. https://github.com/shauli-ravfogel/conformal-prediction.

UR - http://www.scopus.com/inward/record.url?scp=85167956384&partnerID=8YFLogxK

U2 - 10.18653/v1/2023.findings-acl.3

DO - 10.18653/v1/2023.findings-acl.3

M3 - ???researchoutput.researchoutputtypes.contributiontobookanthology.conference???

AN - SCOPUS:85167956384

T3 - Proceedings of the Annual Meeting of the Association for Computational Linguistics

SP - 27

EP - 34

BT - Findings of the Association for Computational Linguistics, ACL 2023

PB - Association for Computational Linguistics (ACL)

T2 - 61st Annual Meeting of the Association for Computational Linguistics, ACL 2023

Y2 - 9 July 2023 through 14 July 2023

ER -

Conformal Nucleus Sampling

Abstract

Publication series

Conference

Bibliographical note

Funding

Access to Document

Other files and links

Fingerprint

Cite this