Concept · How AI answers work
Grounding
Definition
Tying an answer to retrieved sources so its claims can be traced rather than invented.
What it means in practice
A grounded answer carries citations the reader can open. Grounding is not a guarantee of accuracy: a source can be cited and still be misread, and some engines attach sources after composing the text, so a citation is evidence the document was retrieved rather than proof it was used. This is why a mention and a citation have to be counted separately.
How refd treats it
This is category vocabulary, not a refd metric. refd measures what AI answers actually said across repeated runs, and publishes the raw answer behind every number. Where a term above carries a claim about what causes visibility, refd treats that claim as a hypothesis to test rather than a setting to configure.
This definition is editorial. It describes how the term is used across the field rather than something the software computes, so unlike the metric pages it is not generated from the product.
See your evidence