ALEXANDRIA, Va., May 12 -- United States Patent no. 12,626,691, issued on May 12, was assigned to AMAZON TECHNOLOGIES INC. (Seattle).
"Language model hallucination mitigation using contrastive decoding" was invented by Emilio Fabio Monti (Cambridge, Great Britain), Jens Lehmann (Dresden, Germany), Hitham Ahmed Assem Aly Salama (Cambridge, Great Britain) and Zheng Zhao (Edinburgh, Great Britain).
According to the abstract* released by the U.S. Patent & Trademark Office: "Devices and techniques are generally described to mitigate hallucination for language models (LMs) using contrastive decoding. In various examples, first context data and first adversarial data may be determined based on a natural language input. First prompt including the...