arXiv - CSCL: "Can Pretrained Language Models Derive Correct Sem…" - Qoto Mastodon

arXiv - CSCL @arxiv_cscl@qoto.org

Can Pretrained Language Models Derive Correct Semantics from Corrupt Subwords under Noise?. (arXiv:2306.15268v1 [cs.CL])

http://arxiv.org/abs/2306.15268 #arXiv #NLProc

Jun 28, 2023, 03:29 · · arxiv-cscl · · ·

Sign in to participate in the conversation