US Patent Issued to Palo Alto Networks on Sept. 22 for "Automated LLM data leakage detection via persuasive prompting" (California, Washington Inventors)

ALEXANDRIA, Va., Sept. 24 -- United States Patent no. 12,743,526, issued on Sept. 22, was assigned to Palo Alto Networks Inc. (Santa Clara, Calif.).

"Automated LLM data leakage detection via persuasive prompting" was invented by Feng Xiao (San Jose, Calif.), Yang Ji (San Jose, Calif.), Wenjun Hu (Santa Clara, Calif.), Danny Tsechansky (Kirkland, Wash.) and Ali Islam (Fremont, Calif.).

According to the abstract* released by the U.S. Patent & Trademark Office: "Assessments of guardrails of LLMs, whether used by an application or within an AI/LM stack, must be dynamic to protect against the ongoing engineering of jailbreaking prompts. An assessment framework has been created that facilitates assessment of language model guardrails. The asses...