TigreGotico/bifonia-pt-homographs-wild
bifonia-pt-homographs-wild — real-world OOD evaluation set Real Portuguese sentences containing 112 heterophonic homographs — words spelled identically whose pronunciation (IPA) depends on meaning, e.g. sede (thirst ˈsedɨ vs seat ˈsɛdɨ, both nouns), forma (mould ˈfoɾmɐ vs shape ˈfɔɾmɐ), molho (sauce ˈmoʎu vs bundle ˈmɔʎu). Unlike the synthetic TigreGotico/bifonia-pt-homographs train/test set (LLM-generated), every sentence here is mined from real-world text. It is meant as an… See the full description on the dataset page: https://huggingface.co/datasets/TigreGotico/bifonia-pt-homographs-wild.
full per-label verification applied
fix NOUN/VERB mislabels (troco/etc.)
full validation: re-derive pos+ipa, dedup
normalize IPA glyph to ɡ
re-key on meaning: relabel clobber-prone sentences to fine senses (court_of_law/court_martial/manner/currency/around…), drop toponyms
Upload README.md with huggingface_hub
Upload test.jsonl with huggingface_hub
Real-world OOD eval set: Wikipedia + cc100, Sonnet-annotated
initial commit
