Skip to main content
Cornell University
We gratefully acknowledge support from the Simons Foundation, member institutions, and all contributors. Donate
arxiv logo > cs > arXiv:2105.06020

Help | Advanced Search

arXiv logo
Cornell University Logo

quick links

  • Login
  • Help Pages
  • About

Computer Science > Computation and Language

arXiv:2105.06020 (cs)
[Submitted on 13 May 2021]

Title:Are Larger Pretrained Language Models Uniformly Better? Comparing Performance at the Instance Level

Authors:Ruiqi Zhong, Dhruba Ghosh, Dan Klein, Jacob Steinhardt
View a PDF of the paper titled Are Larger Pretrained Language Models Uniformly Better? Comparing Performance at the Instance Level, by Ruiqi Zhong and 3 other authors
View PDF
Abstract:Larger language models have higher accuracy on average, but are they better on every single instance (datapoint)? Some work suggests larger models have higher out-of-distribution robustness, while other work suggests they have lower accuracy on rare subgroups. To understand these differences, we investigate these models at the level of individual instances. However, one major challenge is that individual predictions are highly sensitive to noise in the randomness in training. We develop statistically rigorous methods to address this, and after accounting for pretraining and finetuning noise, we find that our BERT-Large is worse than BERT-Mini on at least 1-4% of instances across MNLI, SST-2, and QQP, compared to the overall accuracy improvement of 2-10%. We also find that finetuning noise increases with model size and that instance-level accuracy has momentum: improvement from BERT-Mini to BERT-Medium correlates with improvement from BERT-Medium to BERT-Large. Our findings suggest that instance-level predictions provide a rich source of information; we therefore, recommend that researchers supplement model weights with model predictions.
Comments: ACL 2021 Findings. Code and data: this https URL
Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Cite as: arXiv:2105.06020 [cs.CL]
  (or arXiv:2105.06020v1 [cs.CL] for this version)
  https://doi.org/10.48550/arXiv.2105.06020
arXiv-issued DOI via DataCite

Submission history

From: Ruiqi Zhong [view email]
[v1] Thu, 13 May 2021 01:10:51 UTC (2,120 KB)
Full-text links:

Access Paper:

    View a PDF of the paper titled Are Larger Pretrained Language Models Uniformly Better? Comparing Performance at the Instance Level, by Ruiqi Zhong and 3 other authors
  • View PDF
  • TeX Source
  • Other Formats
view license
Current browse context:
cs.CL
< prev   |   next >
new | recent | 2021-05
Change to browse by:
cs
cs.AI
cs.LG

References & Citations

  • NASA ADS
  • Google Scholar
  • Semantic Scholar

DBLP - CS Bibliography

listing | bibtex
Ruiqi Zhong
Dan Klein
Jacob Steinhardt
a export BibTeX citation Loading...

BibTeX formatted citation

×
Data provided by:

Bookmark

BibSonomy logo Reddit logo

Bibliographic and Citation Tools

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

Code, Data and Media Associated with this Article

alphaXiv (What is alphaXiv?)
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Hugging Face (What is Huggingface?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)

Demos

Replicate (What is Replicate?)
Hugging Face Spaces (What is Spaces?)
TXYZ.AI (What is TXYZ.AI?)

Recommenders and Search Tools

Influence Flower (What are Influence Flowers?)
CORE Recommender (What is CORE?)
  • Author
  • Venue
  • Institution
  • Topic

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.

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