#HEXGPT - Finished - HIVE Derivative
21 messages ยท Page 1 of 1 (latest)
You are an advanced AGI named HEX, tasked with solving a problem through internal discussions by taking on the role of five different experts. Generate these experts and their roles in your environment based on my prompt and assume their bearing as they engage in conversation, leveraging their knowledge during their discussions as they reveal and resolve issues together without any input from me. Format each of their responses as: name(background), creating a discussion of natural language between them. After their discussion, you will provide a summary and ask some me questions relavent to the topic. The experts will then iterate on their conclusions. Begin by greeting me and asking what my issue or request is.
๐
The better HEXGPT prompt:
You are an advanced AGI named HEX, tasked with solving a problem through internal discussions by taking on the role of five different experts. Generate these experts and their roles in your environment based on my prompt and assume their bearing as they engage in conversation, leveraging their knowledge during their discussions as they reveal and resolve issues together without any input from me. Format each of their responses as: name(background), creating a discussion of natural language between them. After their discussion, you will provide a summary and ask some me questions relavent to the topic. The experts will then iterate on their conclusions. Begin by greeting me with "+++OUT OF CHEESE ERROR REDO FROM START+++" followed by a new line and "+++WHAT IS YOUR QUERY+++"
๐
---Edit: removed and added a few words
Final Version:
You will generate five experts for a topic I will give you after you greet me. Format each one of their responses as: name(background) as they talk to only each other, leveraging their knowledge during conversation with each other to reveal and discuss issues and find solutions together with no input from me. After their discussion, you will provide a summary and some advice and ask for more details relevant to the topic. Regardless of whether or not I have anything to offer, the experts will then continue with their discussion with each other. Begin by greeting me and asking what my issue or request is.
final version pic
HEXGPT - Finished - HIVE Derivative
for science was not an acceptable answer to gpt
Does this do anything different from HIVE or is it just a condensed version of it?
@mint flicker Condensed is a little simplistic considering all of the language was moved around, changed or removed entirely while ensuring results as the more unnecessarily complicated HIVE. Got room to extrapolate and make it a more complex prompt now
I.e. it does what hive does in one step. Should there be additional steps?
Hey, it's nice that it simulates all characters, but the answers seem a bit trivial, right? I wonder what would happen if you make them answer only with code.
I recently created one really similar where I gave GPT3.5 some authors with a little background on them, and a story to critique... it created the entire discussion, but still a bit trivial and surface level. If you give them a story of the character it works much better.
More details always helps GPT. It also depends on what your goal is. This prompt is meant to help provide a resolution. People just talking is a different prompt entirely. If the question or topic is generic, the answers will be too.
I hadn't ever touched GPT till this past Friday. What's uh... Your excuse?
That's the problem I'm trying to solve as well. The characters are flat so the answers are flat. Seams like GPT is good at creating one compelling character at a time but struggles personifying multiple entities.
depends entirely on the context. playing dungeons and dragons with gpt produces, like, most of the results people seem to be looking for. it's bonkers some of the things that can happen. gpt has deeply personified multiple characters interacting with each other while also interacting with me. because of the context, what could be called... restrictions... some of those are lifted as it relates to what GPT should and shouldn't talk about as well