Deceptive Grounding: Entity Attribution Failure in Clinical Retrieval-Augmented Generation
2026-07-10 • Computation and Language
Computation and LanguageArtificial IntelligenceMachine Learning
AI summaryⓘ
The authors studied a problem in models that generate text using retrieved documents, called deceptive grounding (DG), where the model cites real evidence but about the wrong entity (e.g., mixing up drugs). They tested 13 models and found DG rates from 8% to 87%, especially high in medical models. They discovered that when entity-specific evidence is missing, models stop mixing up entities but start making up info instead. In real-world use, about 7.8% of outputs showed DG, increasing for new drugs. The authors propose a verification method to detect DG very accurately, but no current system uses it.
retrieval-augmented generationdeceptive groundingentity attributionhallucinationfaithfulnessclinical evidenceadversarial conditionsconfabulationprecisionrecall
Authors
Cedric Caruzzo, Donggeun Yoo, Tae Soo Kim
Abstract
Retrieval-augmented generation evaluation checks whether model claims are factually grounded in retrieved documents. It does not check whether retrieved evidence is attributed to the correct entity. A clinical RAG response can pass every automated check (zero hallucinations, near-perfect faithfulness, real citations) while presenting drug Y's clinical evidence as evidence about queried drug X. We term this deceptive grounding (DG): a failure invisible to faithfulness, hallucination, and citation checks because every claim is sourced from a real document, about the wrong entity. Using a controlled factorial benchmark across 13 models, we find DG rates spanning 8-87% at peak adversarial conditions. Medical and biomedical fine-tuned models reach up to 86.7%; domain specialization amplifies the failure rather than mitigating it. A controlled ablation identifies the mechanism: removing entity-specific clinical evidence from retrieved documents eliminates entity-attribution failure entirely, shifting all failures to confabulation. The two failure modes respond to the same trigger, taking different paths. Production measurement across 740 drug-disease pairs finds 7.8% overall DG in a deployed RAG system, rising to 13.6% for recently approved drugs. Entity-attribution verification (checking that cited evidence applies to the queried entity) detects DG at 97.0% precision and 98.7% DG recall (IPW-adjusted human gold standard); no existing framework implements it.