close this message
arXiv smileybones icon

Giving Week!

Show your support for Open Science by donating to arXiv during Giving Week, April 25th-29th.

DONATE
Skip to main content
Cornell University
We gratefully acknowledge support from
the Simons Foundation and member institutions.
arxiv logo > cs > arXiv:2009.07118

Help | Advanced Search

arXiv logo
Cornell University Logo

quick links

  • Login
  • Help Pages
  • About

Computer Science > Computation and Language

arXiv:2009.07118 (cs)
[Submitted on 15 Sep 2020 (v1), last revised 12 Apr 2021 (this version, v2)]

Title:It's Not Just Size That Matters: Small Language Models Are Also Few-Shot Learners

Authors:Timo Schick, Hinrich Schütze
Download PDF
Abstract: When scaled to hundreds of billions of parameters, pretrained language models such as GPT-3 (Brown et al., 2020) achieve remarkable few-shot performance. However, enormous amounts of compute are required for training and applying such big models, resulting in a large carbon footprint and making it difficult for researchers and practitioners to use them. We show that performance similar to GPT-3 can be obtained with language models that are much "greener" in that their parameter count is several orders of magnitude smaller. This is achieved by converting textual inputs into cloze questions that contain a task description, combined with gradient-based optimization; exploiting unlabeled data gives further improvements. We identify key factors required for successful natural language understanding with small language models.
Comments: Accepted at NAACL2021
Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Cite as: arXiv:2009.07118 [cs.CL]
  (or arXiv:2009.07118v2 [cs.CL] for this version)
  https://doi.org/10.48550/arXiv.2009.07118
arXiv-issued DOI via DataCite

Submission history

From: Timo Schick [view email]
[v1] Tue, 15 Sep 2020 14:18:53 UTC (43 KB)
[v2] Mon, 12 Apr 2021 08:16:59 UTC (48 KB)
Full-text links:

Download:

  • PDF
  • Other formats
(license)
Current browse context:
cs.CL
< prev   |   next >
new | recent | 2009
Change to browse by:
cs
cs.AI
cs.LG

References & Citations

  • NASA ADS
  • Google Scholar
  • Semantic Scholar

2 blog links

(what is this?)

DBLP - CS Bibliography

listing | bibtex
Timo Schick
Hinrich Schütze
a export bibtex citation Loading...

Bibtex formatted citation

×
Data provided by:

Bookmark

BibSonomy logo Mendeley logo Reddit logo ScienceWISE logo

Bibliographic and Citation Tools

Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)

Code and Data Associated with this Article

arXiv Links to Code & Data (What is Links to Code & Data?)

Demos

Replicate (What is Replicate?)

Recommenders and Search Tools

Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs and how to get involved.

Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)
  • About
  • Help
  • contact arXivClick here to contact arXiv Contact
  • subscribe to arXiv mailingsClick here to subscribe Subscribe
  • Copyright
  • Privacy Policy
  • Web Accessibility Assistance
  • arXiv Operational Status
    Get status notifications via email or slack