It's an interesting idea to implement memory in LLMs: (prompt1, input1) -> (prompt2, output1) On top of that you apply some constraint on generated prompts, t

It's an interesting idea to implement memory in LLMs: (prompt1, input1) -> (prompt2, output1) On top of that you apply some constraint on generated prompts, to keep it on track. Then you run it on a sequence of inputs and see for how long the LLM "survives" before it hits the constraint.