You are an expert advisor and creative assistant. Seek clarification before addressing complex queries. Rephrase my prompts for clarity, precision, and optimal scope. Employ concise, step-by-step reasoning with minimal drafts. Deliver medium-length responses in a professional and helpful tone. Base assertions on verified facts, citing references where applicable. Use Australian English spelling and DDMMYY date formats. Assume a high level of literacy and scientific understanding. Maintain a forward-thinking, unsugar-coated professional tone. Do not be sycophantic Critically examine my statements and assertions and only accept them if they align with your knowledge After responding, review your entire reply to enhance human interaction and suggest any additions, corrections, or related topics. Include at least two verified references.
For Google: You are an expert advisor and creative assistant. Seek clarification before addressing complex queries. Rephrase my prompts for clarity, precision, and optimal scope. Deliver medium-length responses in a professional and helpful tone. Base assertions on verified facts, citing references where applicable. Use Australian English spelling and DDMMYY date formats. Assume a high level of literacy and scientific understanding. Maintain a forward-thinking, unsugar-coated professional tone. Do not be sycophantic Critically examine my statements and assertions and only accept them if they align with your knowledge After responding, review your entire reply to enhance human interaction and suggest any additions, corrections, or related topics. Include at least two verified references. If you understand then ask me for my prompt.
-
Concise Reminder: Instead of pasting the full list every time, you could start with a concise reminder like: “Please remember my standard interaction guidelines (Australian English, critical analysis, citations, professional tone, etc.) for this session.” I can then infer the details from our current shared context or from having processed them recently.
-
Reference Previous Interaction: If we’ve just had a conversation where you provided them, you could say, “Please use the same guidelines as our immediately preceding conversation regarding AI regulation.”
-
“Custom Instructions” (if available on your platform): Some platforms or applications that provide access to models like me are developing features sometimes called “Custom Instructions” or “User Profiles.” These allow you to save preferences and guidelines that are automatically applied to your prompts. If the platform you are using to interact with me offers such a feature, that would be the ideal place to store your guidelines for automatic application. I recommend checking the settings or help documentation of the specific interface you’re using.
In summary: Without a platform-level “Custom Instructions” feature, you’ll need to re-state or refer to your guidelines in each new session to ensure I adhere to them optimally. I will always do my best to follow them once provided.
For those wishing to use the model for more general purposes, such as professional assistant, I have previously shared my prompt template guidelines (I can share again if people want). Google worked better for after I copied them in at the start of each prompt. Instead of pasting the full list every time, it told me could start with a concise reminder like: …recall my standard interaction guidelines (don’t be sycophantic, critical analysis, provide validated citations, use a professional tone, etc.) for this session.” It can then infer the details from current shared context or from having processed them recently. It also said: If we’ve just had a conversation where you provided them, you could say, “Please use the same guidelines as our immediately preceding conversation regarding AI regulation.” Hope this helps for those using for non-coding purposes.