How does the "Calculate similarity score" work under the hood?
The "Calculate similarity score" feature uses our natural language processing algorithms to compare two text inputs and provide a quantitative similarity percentage. You input two pieces of text, and the integration returns a numerical similarity score. Are you looking to understand the specific algorithms or mathematical methods used in the comparison, or do you need help implementing this feature in your workflow?
Does it use embeddings?
The documentation for our "Calculate similarity score" feature doesn't specify the technical implementation details, including whether it uses embeddings. The feature leverages our natural language processing algorithms to compare two text inputs and returns a numerical similarity score, but the specific methods aren't detailed. Are you looking to understand this for a specific use case or integration requirement?
Can you or person find out?
The documentation for our "Calculate similarity score" feature doesn't include technical implementation details like whether it uses embeddings or the specific algorithms involved. The feature leverages natural language processing algorithms to compare two text inputs and returns a numerical similarity score, but the underlying technical methods aren't specified. If you need further technical details about the implementation, you can request human assistance by clicking on "Talk to a person" to connect with our team who may have access to more detailed technical information. What specific use case are you planning to implement this feature for?
Hi Mathew, thanks for reaching out to support!
Are you looking to understand if "Calculate similarity score" uses embeddings?
yes
and if so what algo
Trying to figure out if i need to build my own to improve results or if I'll end up just recreating the same thing
Yes, this functionality does use embeddings and industry-standard similarity algorithms to provide accurate semantic matching between text inputs. Based on what you're trying to accomplish, our current implementation may be exactly what you need and could save you significant development time.