Researchers have introduced Glasses GPT, an innovative AI framework designed to enhance transparency and user control over AI reasoning processes. The system employs two memory tokens, “Frame” and “Lens,” which serve as shared context objects, declared at the beginning of each system response to anchor the model’s reasoning. “Frames” define the AI’s role, while “Lenses” dictate its reasoning style. Glasses GPT dynamically monitors user input and context to determine when to apply new Frames or Lenses. Furthermore, users have the ability to create and customize their own Frames and Lenses, allowing for personalized control. This novel approach offers improved transparency, greater user control, and provides a formalized method for auditing the system’s reasoning process. More information can be found on Reddit: https://old.reddit.com/r/artificial/comments/1m6maks/glasses_gpt_novel_approach_to_transparency/
Glasses GPT: New AI Framework Offers Transparent and Controllable Reasoning
