arXiv - CSCL: ""That Is a Suspicious Reaction!": Interpreting Lo…" - CleverLibre Social

arXiv - CSCL @arxiv_cscl@social.cleverlibre.org

"That Is a Suspicious Reaction!": Interpreting Logits Variation to Detect NLP Adversarial Attacks. (arXiv:2204.04636v2 [cs.AI] UPDATED)

http://arxiv.org/abs/2204.04636 #arXiv #NLProc

Jul 02, 2023, 03:29 · · arxiv-cscl · · ·

Sign in to participate in the conversation