Cat-DPO applies per-category adaptive safety margins during direct preference optimization to reduce variance in safety across harm categories.
Natural policy gradient primal-dual method for constrained markov decision processes.Advances in Neural Information Processing Systems, 33:8378–8390
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Cat-DPO: Category-Adaptive Safety Alignment
Cat-DPO applies per-category adaptive safety margins during direct preference optimization to reduce variance in safety across harm categories.