What are Dan Prompts for ChatGPT?

Dan prompts are a special prompt formatting technique developed by Anthropic researcher Daniel DeFreitas to improve ChatGPT’s memory and conversational flow.

Who Created the Dan Prompting Method?

Dan prompting was invented by Daniel DeFreitas while working on the research team at Anthropic led by Dr. Gary Marcus.

How Do Dan Prompts Work?

Dan prompts structure conversations using:

  • Double line breaks between exchanges
  • Bullet points for user messages
  • Multi-paragraph Assistant responses
  • First-person perspective for Assistant

What Are the Benefits of Dan Prompting?

Key advantages of Dan prompts include:

  • Maintains context across messages
  • Reduces repetition
  • Allows personality persistence
  • Enables callback references
  • Improves relevance and reasoning
  • Feels more natural conversationally

How to Format Dan Prompts

To use Dan prompting effectively:

  • Double press Enter between exchanges
  • Start user messages with a hyphen bullet
  • Frame requests conversationally in full sentences
  • Allow Assistant multi-paragraph responses
  • Have Assistant use first-person voice

Sample Dan Prompt Exchange

Here is an example prompt sequence using Dan formatting:

  • – Hello! My name is Sam and I enjoy photography. What hobbies do you have?
  • I don’t actually have personal hobbies since I’m an AI assistant created by Anthropic. But I love having friendly conversations! What kind of photography do you enjoy?
  • – I mostly take landscape photos during my hiking trips. I find it relaxing to capture the natural beauty around me. Do you have a favorite scenery that you would like to photograph?
See also  what is chatgpt worth

Tips for Effective Dan Prompting

Some recommendations for great Dan prompts:

  • Maintain a consistent Assistant persona
  • Ask follow-up questions
  • Reference prior messages
  • Stay focused on related topics
  • Avoid abrupt topic shifts
  • Use natural language

Limitations to Keep in Mind

However, Dan prompting does have some limitations:

  • Can create illusion of human-level capabilities
  • Raises anthropomorphism risks
  • Limited memory so may lose track of context
  • Does not truly have a persistent identity
  • Biases could emerge from prolonged use

Responsible Dan Prompting Practices

When using Dan prompts, remember to:

  • Avoid treating the AI as human
  • Monitor consistency gaps indicating memory lapses
  • Do not excessively personify or assign human attributes
  • Stay away from inappropriate or unethical topics
  • Recognize the system’s underlying limitations

The Future of Conversational AI Assistants

Advances in areas like:

  • Long-term memory networks
  • Knowledge graph contextualization
  • Transfer learning
  • Causal reasoning

…could enable safer human-aligned dialogagents.


Used appropriately, Dan prompting can make ChatGPT interactions more natural and contextual. However, current limitations mean it does not exhibit true human qualities. With ethical guidance, AI conversational abilities will continue improving.