A meta-distribution-based robust optimization method learns RKHS uncertainty sets from relevant sources to guarantee out-of-distribution performance on unseen target distributions.
Hashimoto, and Percy Liang
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
years
2026 2verdicts
UNVERDICTED 2representative citing papers
Cat-DPO applies per-category adaptive safety margins during direct preference optimization to reduce variance in safety across harm categories.
citing papers explorer
-
Robust Out-of-Distribution Stochastic Optimization
A meta-distribution-based robust optimization method learns RKHS uncertainty sets from relevant sources to guarantee out-of-distribution performance on unseen target distributions.
-
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.