Per-user LLM control layer: personality tuning + moderation explainability

@sama Yeah I quit using ChatGPT because it kept replying both randomly and indiscriminately that my prompt violated some unknown content policy. Some times i would get the bot to generate the content I needed and then I would ask for a moderate tweak and it would say that I violated some unknown content policy. Other times I would get denied a request due to content restrictions and then I would simply reword the prompt and get the exact content I was looking for. Whenever I asked for an explanation of what content I violated, I always received a vague explanation. I don’t need a service to censor my thoughts and requests.