Dataset · tsivakar
Contextual usage examples of ambiguous English words
tsivakar/polysemous-words
A massive web text corpus for lexical sample word sense disambiguation (WSD) and contextual understanding evaluation in language models.
What's inside
It contains usage examples of 200 common polysemous words in natural web texts. The dataset size ranges from 100 million to 1 billion records in English. Each word is presented in multiple different meanings and diverse contexts.
What you can do with it
- Training models for word sense disambiguation and induction (WSD and WSI)
- Testing LLM capability to distinguish meanings of homonyms and polysemous words
- Fine-tuning encoders to improve semantic embedding quality
- Languages
- English
- Size
- 100M–1B rows
- Downloads this month
- 28.7K
- Likes
- 1
How to load it
from datasets import load_dataset
ds = load_dataset("tsivakar/polysemous-words")