arXiv - CSCL: "How Johnny Can Persuade LLMs to Jailbreak Them: R…" - CleverLibre Social

arXiv - CSCL @arxiv_cscl@social.cleverlibre.org

How Johnny Can Persuade LLMs to Jailbreak Them: Rethinking Persuasion to Challenge AI Safety by Humanizing LLMs. (arXiv:2401.06373v2 [cs.CL] UPDATED)

http://arxiv.org/abs/2401.06373 #arXiv #NLProc

Jan 25, 2024, 03:19 · · arxiv-cscl · · ·

Sign in to participate in the conversation