Apple Intelligence's Friendly and Professional text rewrites can substantially reduce LLM-based emotion inference accuracy in small early datasets.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Evaluating Apple Intelligence's Writing Tools for Privacy Against Large Language Model-Based Inference Attacks: Insights from Early Datasets
Apple Intelligence's Friendly and Professional text rewrites can substantially reduce LLM-based emotion inference accuracy in small early datasets.