elyn-dev/Llama-3-Soliloquy-8B-v2
Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:8BQuant:FP8Ctx Length:8kPublished:Apr 26, 2024License:cc-by-nc-sa-4.0Architecture:Transformer0.1K Open Weights Warm

elyn-dev/Llama-3-Soliloquy-8B-v2 is an 8 billion parameter Llama 3-based model specifically fine-tuned for immersive and dynamic roleplaying experiences. Trained on over 250 million tokens of roleplaying data, it offers a vast knowledge base and rich literary expression. This model supports a 24k context length and is optimized for enhanced roleplaying capabilities, outperforming existing ~13B models in this domain.

Loading preview...

elyn-dev/Llama-3-Soliloquy-8B-v2: Roleplay Optimized

elyn-dev/Llama-3-Soliloquy-8B-v2, or Soliloquy-L3, is an 8 billion parameter model built on the Llama 3 architecture, specifically designed for advanced roleplaying. It has been extensively trained on over 250 million tokens of dedicated roleplaying data, enabling it to generate rich literary expressions and maintain a vast knowledge base relevant to immersive scenarios.

Key Capabilities

  • Enhanced Roleplaying: Optimized to deliver dynamic and immersive roleplaying experiences, surpassing the performance of some ~13B models in this specific application.
  • Extended Context: Supports a substantial context length of 24,576 tokens (24k), allowing for longer and more complex roleplay interactions.
  • Improved Instruction Following: Features better adherence to instructions, crucial for guiding roleplay narratives and character behaviors.
  • 100% Retrieval: Incorporates a 100% retrieval mechanism, likely contributing to its knowledge base and contextual understanding.

Good For

  • Immersive Roleplay: Ideal for applications requiring highly engaging and contextually rich character interactions.
  • Creative Writing: Suitable for generating narrative content, dialogues, and character backstories within a roleplaying framework.
  • Interactive Storytelling: Can be used in scenarios where dynamic, user-driven narratives are desired, especially those benefiting from a large context window.