Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Beyond the Imitation Game: Quantifying and extrapolatingthe capabilities of language models
University of Notre Dame, USA.
RISE Research Institutes of Sweden, Digital Systems, Data Science.ORCID iD: 0000-0002-6032-6155
Number of Authors: 4502023 (English)In: Transactions on Machine Learning Research, E-ISSN 2835-8856, no 5Article in journal (Refereed) Published
Abstract [en]

Language models demonstrate both quantitative improvement and new qualitative capabilities with increasing scale. Despite their potentially transformative impact, these new capabilities are as yet poorly characterized. In order to inform future research, prepare for disruptive new model capabilities, and ameliorate socially harmful effects, it is vital that we understand the present and near-future capabilities and limitations of language models. To address this challenge, we introduce the Beyond the Imitation Game benchmark (BIG- bench). BIG-bench currently consists of 204 tasks, contributed by 450 authors across 132 institutions. Task topics are diverse, drawing problems from linguistics, childhood develop- ment, math, common-sense reasoning, biology, physics, social bias, software development, and beyond. BIG-bench focuses on tasks that are believed to be beyond the capabilities of current language models. We evaluate the behavior of OpenAI's GPT models, Google- internal dense transformer architectures, and Switch-style sparse transformers on BIG-bench, across model sizes spanning millions to hundreds of billions of parameters. In addition, a team of human expert raters performed all tasks in order to provide a strong baseline. Findings include: model performance and calibration both improve with scale, but are poor in absolute terms (and when compared with rater performance); performance is remarkably similar across model classes, though with benefits from sparsity; tasks that improve gradually and predictably commonly involve a large knowledge or memorization component, whereas tasks that exhibit "breakthrough" behavior at a critical scale often involve multiple steps or components, or brittle metrics; social bias typically increases with scale in settings with ambiguous context, but this can be improved with prompting.

Place, publisher, year, edition, pages
2023. no 5
National Category
Computer and Information Sciences
Identifiers
URN: urn:nbn:se:ri:diva-67825DOI: 10.48550/arXiv.2206.04615OAI: oai:DiVA.org:ri-67825DiVA, id: diva2:1812528
Available from: 2023-11-16 Created: 2023-11-16 Last updated: 2024-01-08Bibliographically approved

Open Access in DiVA

No full text in DiVA

Other links

Publisher's full textFull text

Authority records

Kleyko, Denis

Search in DiVA

By author/editor
Kleyko, Denis
By organisation
Data Science
In the same journal
Transactions on Machine Learning Research
Computer and Information Sciences

Search outside of DiVA

GoogleGoogle Scholar

doi
urn-nbn

Altmetric score

doi
urn-nbn
Total: 93 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf