SEMANTIC VECTOR SPACE

3D EMBEDDING VISUALIZATION · COSINE SIMILARITY

CORE PRINCIPLE

Word → Vector → Distance

Each word maps to a point in high-dimensional space.
Semantically similar words are closer together.
Model uses cosine similarity to measure relatedness.

similarity = cos(θ)
= (A·B) / (|A|×|B|)
= -1(opposite) ~ +1(same)

INTERACTIVE EXAMPLE

Compare Sentiment Phrases

Select a query phrase to explore:

SIMILARITY SCORES

"I love you"

Cosine similarity with other words:

STEPS

INSTRUCTIONS

Click steps on the left, or click word spheres in 3D space to see similarity scores.

CALCULATION

Select a word to see calculation