Robust Explanations for User Trust in Enterprise NLP Systems
arXiv:2604.12069v1 Announce Type: cross
Abstract: Robust explanations are increasingly required for user trust in enterprise NLP, yet pre-deployment validation is difficult in the common case of black-box deployment (API-only access) where representat…