US Patent Issued to Amazon Technologies on April 14 for "Reward weighting alignment of large language models" (California, Washington, Pennsylvania Inventors)

ALEXANDRIA, Va., April 15 -- United States Patent no. 12,602,548, issued on April 14, was assigned to Amazon Technologies Inc. (Seattle).

"Reward weighting alignment of large language models" was invented by Sailik Sengupta (Santa Clara, Calif.), Daniele Bonadiman (Sunnyvale, Calif.), Yi-An Lai (Bellevue, Wash.), Arshit Gupta (Sunnyvale, Calif.), Dan Roth (Philadelphia), Katrin Kirchhoff (Seattle), Saab Mansour (San Jose, Calif.) and James Yipeng Huang (Los Angeles).

According to the abstract* released by the U.S. Patent & Trademark Office: "This patent application relates to using framework parameters with a large language model to create beams based on a prompt. The beams can be evaluated using multiple criteria of a reward model that c...