Semantic Quoting Query

The 'Sample Content' property should contain an example of the paragraph, clause, or document section you want to detect using semantic search. This text is embedded and compared to segments of the document to find the most similar content.

• Provide a clear and representative sample for best results.
• Use multiple queries if clause wording may vary.

Example:
The parties agree to maintain confidentiality of all information disclosed...

The quality and specificity of the sample directly impact the accuracy of semantic matching.

The 'Token Count' property displays the number of tokens in the 'Sample Content', as determined by the current tokenization method. This is useful for understanding the size and complexity of the sample, and for optimizing chunking and embedding operations.

• Shorter samples may match more broadly, while longer samples provide more specificity.
• Use this value to help tune chunking and model selection for your use case.