File size: 1,567 Bytes
ec76dbf |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 |
base_model: mistralai/Mistral-7B-Instruct-v0.2
gate_mode: hidden # one of "hidden", "cheap_embed", or "random"
dtype: bfloat16 # output dtype (float32, float16, or bfloat16)
experts:
- source_model: maywell/PiVoT-0.1-Starling-LM-RP
positive_prompts:
- "This is an RP session. Respond by guessing the character’s behavior based on user input."
- source_model: SanjiWatsuki/Kunoichi-DPO-v2-7B
positive_prompts:
- "Think of yourself as a novelist and respond by imagining what would follow the user's input."
- source_model: NeverSleep/Noromaid-7B-0.4-DPO
positive_prompts:
- "You are an adult fiction author. Read the user's input and write down the next possible situation."
- source_model: maywell/Synatra-7B-v0.3-RP
positive_prompts:
- "Pretend you are the writer and take the user's input to guess what comes next and respond."
- source_model: berkeley-nest/Starling-LM-7B-alpha
positive_prompts:
- "This is an RP session. Respond by guessing the character’s behavior based on user input."
- source_model: Intel/neural-chat-7b-v3-3
positive_prompts:
- "Think of yourself as a novelist and respond by imagining what would follow the user's input."
- source_model: mistralai/Mistral-7B-v0.1
positive_prompts:
- "You are an adult fiction author. Read the user's input and write down the next possible situation."
- source_model: mlabonne/NeuralBeagle14-7B
positive_prompts:
- "Pretend you are the writer and take the user's input to guess what comes next and respond." |