Same Agent, Different Answers: A Repeat-Aware Audit of Corpus-Induced Answer Churn in Retrieval-Augmented QA
2026-08-24 • Information Retrieval
Information RetrievalComputation and Language
AI summaryⓘ
The authors study how a question-answering system that uses retrieval can give different answers after its search index is expanded, even when all other settings stay the same. They find that changes in answers can be hidden if you only look at overall accuracy, which might look stable. To measure this hidden variability, they introduce a method called the Snapshot Compatibility Audit to detect when answers change more than expected. Their experiments show significant answer changes without big drops in accuracy, meaning the system’s responses can be unstable in subtle ways. They suggest QA system updates should check for answer consistency, not just accuracy improvements.
retrieval-augmented QAindex expansionanswer churnexact-match accuracySnapshot Compatibility AuditNatural QuestionsTriviaQAsemantic flipsrepeat-stability
Authors
Jingjie Ning, Xueqi Li
Abstract
A retrieval-augmented QA system can return different answers after an index expansion even when its requested model identifier, prompt, retrieval policy, evidence depth, rendering, and exposed generation controls are held fixed. Aggregate accuracy may hide these changes when gains and losses cancel, while ordinary generation variability makes one-shot comparisons overstate update effects. We call the hidden phenomenon accuracy-blind answer churn and introduce the \emph{Snapshot Compatibility Audit}, which estimates excess answer churn by subtracting same-snapshot repeat disagreement from cross-snapshot disagreement. We instantiate it by expanding one frozen FineWeb prefix from one to seven shards. In a preregistered 400-question Natural Questions study, normalized-exact and blinded-semantic excess churn are 6.44 and 10.25 percentage points while exact-match accuracy changes by only $-1.50$ points. A post-hoc analysis finds repeat-stable semantic flips on 40/400 questions. A separately preregistered 200-question TriviaQA study yields smaller, directionally consistent excess churn while exact-match accuracy moves in the opposite direction. An outcome-blind post-hoc 100-question subset replication with a second DeepSeek generator and serving configuration finds 8.75 pp of semantic excess churn even as exact match rises by 3.00 percentage points. Answer-level compatibility can therefore fail without a conspicuous or consistently directed utility shift. Retrieval-augmented releases should audit compatibility alongside utility.