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 Plus annual

Presents the access of premium content as premium feature

Premium Content

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 Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Word co-occurrence patterns in language corpora contain a surprising amountof conceptual knowledge. Large language models (LLMs), trained to predict wordsin context, leverage these patterns to achieve impressive performance ondiverse semantic tasks requiring world knowledge. An important but understudiedquestion about LLMs' semantic abilities is whether they acquire generalizedknowledge of common events. Here, we test whether five pre-trained LLMs (from2018's BERT to 2023's MPT) assign higher likelihood to plausible descriptionsof agent-patient interactions than to minimally different implausible versionsof the same event. Using three curated sets of minimal sentence pairs (totaln=1,215), we found that pre-trained LLMs possess substantial event knowledge,outperforming other distributional language models. In particular, they almostalways assign higher likelihood to possible vs. impossible events (The teacherbought the laptop vs. The laptop bought the teacher). However, LLMs show lessconsistent preferences for likely vs. unlikely events (The nanny tutored theboy vs. The boy tutored the nanny). In follow-up analyses, we show that (i) LLMscores are driven by both plausibility and surface-level sentence features,(ii) LLM scores generalize well across syntactic variants (active vs. passiveconstructions) but less well across semantic variants (synonymous sentences),(iii) some LLM errors mirror human judgment ambiguity, and (iv) sentenceplausibility serves as an organizing dimension in internal LLM representations.Overall, our results show that important aspects of event knowledge naturallyemerge from distributional linguistic patterns, but also highlight a gapbetween representations of possible/impossible and likely/unlikely events.

Event knowledge in large language models: the gap between the impossible and the unlikely