Citation Grounding
A rigorous generation technique where an AI model explicitly links every factual claim it makes back to a specific sentence in a retrieved source document.
Think of It Like This
Like a lawyer who isn't allowed to just make a claim to the jury, but must physically hold up the exact piece of evidence that proves their statement.
In high-stakes Retrieval-Augmented Generation (RAG) systems, simply providing the model with context is not enough to stop hallucinations. By forcing the model to emit a citation token for every sentence it generates, developers can programmatically verify that the output perfectly matches the provided evidence. This transforms a black-box generation into an auditable, trustworthy response.