GENEVA, May 19 -- GOOGLE LLC (1600 Amphitheatre ParkwayMountain View, California 94043) filed a patent application (PCT/US2024/054737) for "HYBRID INFERENCE FOR AN EFFICIENT, LOW LATENCY LLM-BASED ASSISTANT" on Nov 06, 2024. With publication no. WO/2025/101616, the details related to the patent application was published on May 15, 2025.

Notably, the patent application was submitted under the International Patent Classification (IPC) system, which is managed by the World Intellectual Property Organization (WIPO).

Inventor(s): SHARIFI, Matthew (Brandschenkestrasse 1108002 Zurich), CARBUNE, Victor (Brandschenkestrasse 1108002 Zurich)

Abstract: Implementations utilize a hybrid use of a smaller LLM and a larger LLM to generate and refine co...