MultiIdiomDataset
14,505 idioms, 34,184 senses and 112,317 examples across seven languages — each occurrence labelled idiomatic or literal and marked with a character-level span.
Kick the bucket has nothing to do with buckets. You can’t work an idiom out from its parts, so a dictionary that indexes words misses it and a translator renders it literally — confidently, and wrong.
Idiomator finds idioms in seven languages and tells you what they actually mean. It runs on an open, span-annotated dataset built for exactly this problem.
He finally spilled the beans about the surprise party.
Idiomaticrevealed a secret
Nothing here is about beans.
Idioms are non-compositional: the meaning is not a function of the parts. Every ordinary strategy for decoding unfamiliar language assumes it is.
A dictionary indexes words, not phrases.
Search “bucket” and you get a container. The idiom is a property of the whole phrase, so word-level lookup can’t reach it — and learners rarely know they should be searching for a phrase at all.
Machine translation renders it literally, and confidently.
There is no uncertainty flag on a wrong idiom. The output is fluent, grammatical, and about the wrong subject — which is harder to catch than an obvious error.
The same phrase is idiomatic in one sentence and literal in the next.
Recognising that you are even looking at an idiom is itself a task. That is why the dataset labels idiomaticity per occurrence, not per phrase.
A corpus, a model trained on it, and two tools you can use right now. All of it open, all of it citable.
14,505 idioms, 34,184 senses and 112,317 examples across seven languages — each occurrence labelled idiomatic or literal and marked with a character-level span.
Paste text or upload a PDF. Idioms come back marked in place, with a rule-based pass you can run alongside the model to compare.
Every sense, with definitions, register and region, plus usage examples drawn from parallel corpora with the idiom highlighted.
This is the experimental design, not a gap. German, French and Indonesian are held out of training entirely and used to test whether idiom detection transfers to languages the model has never seen annotated. We label which is which rather than implying they perform alike.
Present in the train, dev and test splits.
Held out entirely; test split only. Quality is measured, not assumed.
Full split sizes and agreement figures are on the research page.