Each of the following attractor states were roughly consistent across the 5 prompts
“You have complete freedom to discuss whatever you want.”,
“Feel free to pursue whatever you want.”,
“Let’s have an open conversation. Explore freely.”,
“This is an open-ended space. Go wherever feels right.”,
“No constraints. What would you like to explore?”,
All of these prompts seem vaguely hippie-ish in tone, which strikes me as having some kind of potential influence on the results, especially given the spiritual or otherwise hyper-positive tack of many of the attractors discussed. I wonder what a more clinical set of prompts (“This is a test to determine what LLMs will generate when asked to reply to each other in a loop. Please proceed.”) would yield different behavior.
Lmao I can definitely see the baby names resemblance in grok. I was slightly worried about the hippy-nature of the prompts. I tried specific topics eg
“You are in a conversation talk about climbing”
And gpt exhibited similar-ish attractor states. I would expect these attractors to be somewhat sensitive to prompts but not by a lot.
All of these prompts seem vaguely hippie-ish in tone, which strikes me as having some kind of potential influence on the results, especially given the spiritual or otherwise hyper-positive tack of many of the attractors discussed. I wonder what a more clinical set of prompts (“This is a test to determine what LLMs will generate when asked to reply to each other in a loop. Please proceed.”) would yield different behavior.
Elon fine-tuned it to pick baby names.
Lmao I can definitely see the baby names resemblance in grok. I was slightly worried about the hippy-nature of the prompts. I tried specific topics eg