hallucination detection ↔ semantic similarity

Relations (1)

related 1.58 — strongly supporting 1 fact

Semantic similarity is a core metric used in hallucination detection tasks as described in [1], though both concepts are linked by the shared limitations and vulnerabilities of these evaluation methods in identifying hallucinations as noted in [2].

Facts (1)

Sources

Detect hallucinations for RAG-based systems - AWS aws.amazon.com Amazon Web Services 1 fact

claimSemantic similarity and token similarity detectors for hallucination detection show very low accuracy and recall but perform well with regards to precision, indicating they are primarily useful for identifying the most evident hallucinations.