The author introduces FENICE, a novel metric for factuality evaluation in summarization, leveraging NLI-based alignments between claims extracted from the summary and the source text at multiple levels of granularities.
Large language models pose challenges in factual consistency in summarization, addressed by FENICE, a novel metric using NLI and claim extraction.
FENICE proposes a novel factuality evaluation metric for summarization based on NLI and claim extraction, achieving state-of-the-art performance.
FENICE bietet eine effiziente und interpretierbare Metrik zur Bewertung der Faktentreue von Zusammenfassungen.