RAC delivers stable 96% accuracy and up to 94% F1 on unbalanced confidential document classification while keeping sensitive information out of model parameters, matching or exceeding fine-tuned models in robustness.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CR 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Retrieval Augmented Classification for Confidential Documents
RAC delivers stable 96% accuracy and up to 94% F1 on unbalanced confidential document classification while keeping sensitive information out of model parameters, matching or exceeding fine-tuned models in robustness.