Visualize word vectors in 2D (via PCA) and explore analogies like E(queen) - E(king) + E(uncle) ≈ E(aunt).
You can edit the vectors, select words for analogy, and see the result visualized.
| Word | Dim 1 | Dim 2 | Dim 3 | Dim 4 |
|---|---|---|---|---|
| king | ||||
| queen | ||||
| man | ||||
| woman | ||||
| uncle | ||||
| aunt | ||||
| prince | ||||
| princess |
Analogy:-+=
How Word Embeddings Work:
- 1. Words are represented as high-dimensional vectors
- 2. Similar words have similar vector representations
- 3. Analogies can be computed using vector arithmetic
- 4. Example: king - man + woman ≈ queen
- 5. PCA reduces dimensions for 2D visualization
- 6. Cosine similarity measures vector closeness
