Inference Context Stability
Definition:
Inference context stability describes whether a generative system infers the same meaning from a content segment across different prompts, queries, and retrieval contexts. In AI search, unstable inference reduces confidence and leads to suppression even when content is correct and extractable.
Full reference: Inference Context Stability in Generative Search