A RAG assistant started giving wrong citations even though answer quality looks stable. Diagnose it.

Focus on identifying potential issues in the retrieval-augmented generation (RAG) process that could lead to incorrect citations. Consider the various components involved, such as the data sources, retrieval mechanisms, and the model's citation generation logic. Analyze whether the problem lies in the training data, the retrieval algorithm, or the way citations are extracted and formatted. Discuss how you would validate your hypotheses and what metrics you would use to assess the accuracy of citations versus answer quality. Finally, propose actionable steps to investigate and resolve the issue.