US Patent Issued to MCKINSEY & Co. on July 7 for "Contextualized output reliability evaluation for language models" (Arizona, Illinois, Texas Inventors)

ALEXANDRIA, Va., July 7 -- United States Patent no. 12,675,649, issued on July 7, was assigned to MCKINSEY & COMPANY INC. (New York).

"Contextualized output reliability evaluation for language models" was invented by Siddharth Tiwari (Tempe, Ariz.), Francisco Reveriano (Austin, Texas) and Vikramsinh Sohoni (Chicago).

According to the abstract* released by the U.S. Patent & Trademark Office: "A method of evaluating the reliability of a response generated by a language model is described herein. The method includes obtaining a query provided to a language model and a response to the query generated by the language model, and applying hallucination evaluation techniques to one or more of the query and the response to generate hallucination s...