TehVenom/Metharme-13b-Merged

TEXT GENERATIONConcurrency Cost:1Model Size:13BQuant:FP8Ctx Length:4kPublished:May 18, 2023Architecture:Transformer0.0K Cold

TehVenom/Metharme-13b-Merged is a 13 billion parameter instruction-tuned LLaMA model, based on Meta's LLaMA-13b, with a context length of 4096 tokens. Developed by PygmalionAI, this model is specifically biased towards fiction writing, conversation, and roleplaying. It excels at generating narrative content and interactive dialogue, making it suitable for text adventure games and creative writing applications.

Loading preview...

Metharme-13b-Merged Overview

TehVenom/Metharme-13b-Merged is a 13 billion parameter instruction-tuned model derived from Meta's LLaMA-13b, originally developed by PygmalionAI. This model is specifically fine-tuned for conversational AI, roleplaying, and story writing, aiming to provide a model that can be guided by natural language instructions while excelling in creative text generation.

Key Capabilities

  • Instruction-tuned for creative tasks: Optimized for generating fictional narratives, engaging in roleplay scenarios, and handling dynamic conversations.
  • Structured Prompting: Utilizes <|system|>, <|user|>, and <|model|> tokens for clear instruction and turn-based conversation management, allowing for complex interactive experiences like text adventure games.
  • Contextual Understanding: Designed to maintain conversation history and system instructions effectively within its 4096-token context window.

Good For

  • Fictional Writing: Ideal for generating stories, character dialogues, and descriptive passages.
  • Roleplaying: Excels in creating interactive roleplay experiences with guided responses.
  • Conversational Agents: Suitable for building chatbots that require a creative or narrative flair.
  • Text Adventure Games: Demonstrates strong capabilities in describing scenarios and offering choices, as shown in its prompting examples.

Limitations

It is important to note that Metharme-13b-Merged was not fine-tuned for safety or harmlessness. It may produce socially unacceptable, offensive, or factually incorrect content due to its training data, which includes profanity and potentially lewd texts. Its intended use is solely for fictional writing and entertainment purposes.