IdiomX is a large-scale multilingual dataset and benchmark designed to help AI systems understand idiomatic language beyond literal word meanings.
It supports four benchmark tasks:
Idiom Detection
Context → Idiom Retrieval
Arabic → English Idiom Retrieval
Idiom Interpretation… See the full description on the dataset page:
https://huggingface.co/datasets/aymansharara/IdiomX.
