Sandipan99/POLAR
The POLAR Framework: polar Opposites Enable Interpretability of Pre-Trained Word Embeddings
This framework helps researchers and engineers understand what pre-trained word embeddings really mean. It takes existing word embeddings and transforms them into a new format where each dimension is clearly defined by a pair of opposite concepts, like 'cold' vs. 'hot'. The output is more interpretable word embeddings that can be used in various language understanding tasks.
No commits in the last 6 months.
Use this if you need to explain or understand the underlying semantic meaning captured within your pre-trained word embeddings.
Not ideal if you primarily need to improve the performance of your language model without needing to interpret the internal representations.
Stars
13
Forks
8
Language
Roff
License
—
Category
Last pushed
Aug 03, 2020
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/embeddings/Sandipan99/POLAR"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Featured in
Higher-rated alternatives
embeddings-benchmark/mteb
MTEB: Massive Text Embedding Benchmark
harmonydata/harmony
The Harmony Python library: a research tool for psychologists to harmonise data and...
yannvgn/laserembeddings
LASER multilingual sentence embeddings as a pip package
embeddings-benchmark/results
Data for the MTEB leaderboard
Hironsan/awesome-embedding-models
A curated list of awesome embedding models tutorials, projects and communities.