A fact-alignment LLM method detects clinical summarization hallucinations better than existing metrics, with correlations of 0.43 on controlled data and 0.37 on natural errors.
Baseline Methods Our goal is to identify metrics that reliably quantify halluci- nations in clinical text summaries
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
Fact-Controlled Diagnosis of Hallucinations in Medical Text Summarization
A fact-alignment LLM method detects clinical summarization hallucinations better than existing metrics, with correlations of 0.43 on controlled data and 0.37 on natural errors.