Can an LLM lose conceptual continuity while remaining coherent?

Let’s see .. I have been chasing ghosts with my current approach for the TIS training (base system works fine) but I can see some promising signals that could be used for my objectives. It will be nice, though, to try your approach of “offloading” some important concepts to that state and use it as a guideline for the inference pass