RG-TTA uses reinforcement learning at test time to gate fairness regularization by estimated bias sensitivity, reducing stereotypes on FairFace and UTKFace while improving zero-shot utility.
arXiv preprint arXiv:2511.18123 , year =
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.CL 2years
2026 2verdicts
UNVERDICTED 2representative citing papers
A framework estimates grammatical gender directions in contextual embeddings via controlled and natural contexts, finding unweighted controlled contexts and centroid estimators yield the purest directions.
citing papers explorer
-
Selective Test-Time Debiasing for CLIP via Reward Gating
RG-TTA uses reinforcement learning at test time to gate fairness regularization by estimated bias sensitivity, reducing stereotypes on FairFace and UTKFace while improving zero-shot utility.
-
Estimating Grammatical Gender Directions in Contextual Embeddings under Controlled and Natural Contexts
A framework estimates grammatical gender directions in contextual embeddings via controlled and natural contexts, finding unweighted controlled contexts and centroid estimators yield the purest directions.