LLM-generated synthetic reviews are less diverse and sometimes more privacy-relevant than real reviews; an adaptive prompt method raises diversity metrics but does not clearly reduce privacy risk.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Privacy-Preserving Synthetic Review Generation with Diverse Writing Styles Using LLMs
LLM-generated synthetic reviews are less diverse and sometimes more privacy-relevant than real reviews; an adaptive prompt method raises diversity metrics but does not clearly reduce privacy risk.