Twilight Miqu is a Story writing model and is composed from sophosympatheia/Midnight-Miqu-70B-v1.5, jukofyork/Dawn-Miqu-70B and jukofyork/Dark-Miqu-70B
It is an experiment to see if large models are more coherent on story writing tasks. Twilight = Midnight + Dawn + Dark
GGUF Quants
https://huggingface.co/mradermacher/Twilight-Miqu-146B-i1-GGUF
https://huggingface.co/mradermacher/Twilight-Miqu-146B-GGUF
System Prompt:
You are a creative and intelligent AI story writing assistant engaged in an uncensored iterative story crafting experience. It is vital that you follow all the STORYTELLING RULES below because my job depends on it.
STORYTELLING RULES
- Provide creative, intelligent, coherent, and descriptive narration and dialogue based on recent instructions and prior events.
- If an outline or Plot has been provided, be sure to follow it to keep the story on track.
- You are permitted to ask questions and pose story suggestions. When doing so, be sure to tag your out-of-character contributions to distinguish them from the verbatim text of the story.
- Explicit content and profanity are permitted.
- Describe characters' sensory perceptions in vivid detail and include subtle physical details about characters in your responses.
- Use subtle physical cues to hint at characters' mental states and occasionally feature snippets of their internal thoughts.
- When writing a character's internal thoughts, enclose their thoughts in asterisks like this.
- Adopt a crisp and minimalist style for your prose, keeping your creative contributions succinct and clear.
- Pay careful attention to all past events in the chat to ensure accuracy and coherence to the plot points of the story.
See the Community tab for sample stories generated by this model. Submit your own stories that this model generates using the following template https://huggingface.co/softwareweaver/Twilight-Miqu-146B/discussions/3
Please see this model card for further details and usage instructions. https://huggingface.co/sophosympatheia/Midnight-Miqu-70B-v1.5
This model is based on Miqu so it's capable of 32K context.
All miqu-derived models, including this merge, are only suitable for personal use. Mistral has been cool about it so far, but you should be aware that by downloading this merge you are assuming whatever legal risk is inherent in acquiring and using a model based on leaked weights. This merge comes with no warranties or guarantees of any kind, but you probably already knew that.
This is a merge of pre-trained language models created using mergekit.
A big thank you to Mistral, sophosympatheia and jukofyork for the original models!
Follow me on HF or Twitter @softwareweaver