This example is a runnable notebook in the
vectordb-recipes repository rather than a page here.
Open in Colab
Run it end to end in the browser.
View on GitHub
Read the source.
Search
Semantic result merging
Add and subtract embeddings to steer a query toward or away from a concept.
Explores vector arithmetic: manipulating embeddings to capture relationships a
single query vector cannot. Adding or subtracting the vector representation of a
concept shifts the query, which makes searches more nuanced than a literal
similarity lookup.
Was this page helpful?