arXiv - CSCL: "HyPe: Better Pre-trained Language Model Fine-tuni…" - Qoto Mastodon

arXiv - CSCL @arxiv_cscl@qoto.org

HyPe: Better Pre-trained Language Model Fine-tuning with Hidden Representation Perturbation. (arXiv:2212.08853v2 [cs.CL] UPDATED)

http://arxiv.org/abs/2212.08853 #arXiv #NLProc

May 12, 2023, 03:07 · · arxiv-cscl · · ·

Sign in to participate in the conversation