I mean, most llm makers work pretty hard to conceal the system prompt, and I have no idea why XAi give Grok access to a database of historical prompts. LLMs don’t have memories by default, and their inability to learn from past experiences is kind of a big stumbling point for a lot of folks. You can ask, but I doubt you’re likely to get anything other than a confabulation.