Word Embeddings Playground

Enter words into the formula to explore word relationships. The calculation uses vector arithmetic: Word_A - Word_B + Word_C.

How it works: Each word is represented as a vector of numbers. The formula performs vector arithmetic: it takes the difference between A and B (capturing their relationship), then adds that to C to find a word with a similar relationship to C.

Multiple modes: Leave fields empty to try different operations:
• All fields: A - B + C (traditional analogy)
• Just A and B: A - B (relationship difference)
• Just A and C: A + C (concept combination)
• Just A: Find similar words

Choose embedding dataset size:

Current: 1,000 words (embeddings-1k.json)
Dimensions: 300
Status: Loading...

- + = ?

Example: king - man + woman → expected: queen

Try different modes: • Clear Word C for: king - man = ? • Clear Word B for: king + woman = ? • Clear B&C for: similar to king

Results:

Enter words above and click "Calculate Analogy".

Input Word Embeddings:

Word A ():

Word B ():

Word C ():

Formula Result Vector:

Result of A - B + C:

Distance Analysis:

Understanding Similarity Metrics:

Cosine Similarity: Measures the angle between vectors (0 to 1, higher = more similar)
Euclidean Distance: Measures straight-line distance between vectors (lower = more similar)

Click on any result word to see distance calculations.

Available Words Preview:

Sample words available: Loading...

Total words loaded: 0