CARE uses guard-model detection, token-buffer rollback, and self-reflective prompting to reduce harmful responses while preserving response quality.
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
-
CARE: Decoding Time Safety Alignment via Rollback and Introspection Intervention
CARE uses guard-model detection, token-buffer rollback, and self-reflective prompting to reduce harmful responses while preserving response quality.