Swarm intelligence without degradation using two Qwen models

I recently started a weekend project—partly because I was thinking about cancelling my AI subscriptions—with the goal of creating a system capable of handling long contexts, maintaining persistent memory, and engaging in deep reasoning. To achieve this, I decided to modify a model by structuring its data handling across six axes (details are available on GitHub). In short, by adopting this structure and eliminating unnecessary complexity, the model exchanges information via vectors rather than through natural language—a shift away from the "swarm intelligence" approach used previously. This Space separates the functions into a "thinking model" and a "translating model." Please try entering something into the Space via the link provided. Currently, having lost its spatial anchors, the system has effectively become a machine that outputs philosophical musings. If anyone is able to analyze this model, I would appreciate your insights on how to get it to output coherent, normal text. A key strength of this model—and of "jgen"—is its ability to visualize what the AI is thinking in real-time. https://huggingface.co/spaces/kofdai/Verantyx-God-Mode