English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

When people think of everyday things like an egg, they typically have amental image associated with it. This allows them to correctly judge, forexample, that "the yolk surrounds the shell" is a false statement. Do languagemodels similarly have a coherent picture of such everyday things? Toinvestigate this, we propose a benchmark dataset consisting of 100 everydaythings, their parts, and the relationships between these parts, expressed as11,720 "X relation Y?" true/false questions. Using these questions as probes,we observe that state-of-the-art pre-trained language models (LMs) like GPT-3and Macaw have fragments of knowledge about these everyday things, but do nothave fully coherent "parts mental models" (54-59% accurate, 19-43% conditionalconstraint violation). We propose an extension where we add a constraintsatisfaction layer on top of the LM's raw predictions to apply commonsenseconstraints. As well as removing inconsistencies, we find that this alsosignificantly improves accuracy (by 16-20%), suggesting how the incoherence ofthe LM's pictures of everyday things can be significantly reduced.

Do language models have coherent mental models of everyday things?