1 paper · 2 filters
Ashim Gupta, Rishanth Rajendhran, Nathan Stringham +2
Do larger and more performant models resolve NLP's longstanding robustness issues? We investigate this question using over 20 models of different sizes spanning different architect…