Open-AI is developing new reasoning model 'Strawberry' to enhance Large Language Models

From Parhlo Media: 2024-07-15 08:32:33

Developers of Chat-GPT Open-AI are creating a new reasoning model called ‘Strawberry’ to enhance their Large Language Models (LLMs). ‘Strawberry’ will help LLMs solve complex problems and plan ahead, with the Long Horizon Task (LHT) being a key part of this endeavor.

Originally a secret project named Q*, Open-Ai’s new LLM model aims to solve previously unsolved math problems and enable autonomous internet navigation. This will allow Open-Ai to conduct deep research and expand their capabilities significantly.

The new model will utilize large-based data to post-train Open-AI, enabling it to plan ahead and navigate in specific ways even with generalized data. ‘Strawberry’ will be used for the Long Horizon Task (LHT), with computer-using agents (CUA) supporting autonomous search to ensure quick and efficient responses from Chat-GPT.



Read more at Parhlo Media: Open-AI is Developing a New Reasoning Technology Named ‘Strawberry’