Skip to content
conftrace_

intended feature recovery

1 paper

Explore in graph

Co-occurring keywords · papers mentioning both

model robustification (1) feature subspace (1) bias mitigation (1) out-of-distribution generalization (1) natural language understanding (1) model robustness (1)

Papers

Less Is Better: Recovering Intended-Feature Subspace to Robustify NLU Models COLING 2022