Pygmalion-2-13B
| Property | Value |
|---|---|
| Parameter Count | 13B |
| Framework | PyTorch |
| License | Llama2 |
| Tensor Type | BF16 |
What is pygmalion-2-13B?
Pygmalion-2-13B is an advanced language model built on Meta AI's Llama-2 architecture, specifically designed for creative writing and conversational AI applications. This model represents a significant evolution in the Pygmalion series, incorporating instruction-tuning capabilities while maintaining strong performance in fictional content generation and interactive dialogue.
Implementation Details
The model utilizes a sophisticated prompting system with three distinct tokens: <|system|>, <|user|>, and <|model|>. It's trained on multiple datasets including PIPPA, OpenOrca, and other carefully curated sources focused on roleplay and instruction following.
- Built on Llama-2 13B architecture
- Implements specialized token system for different interaction modes
- Trained on five major datasets
- Supports both commercial and non-commercial use
Core Capabilities
- Fiction writing and storytelling
- Natural conversation and roleplay
- Instruction following with contextual awareness
- Multi-turn dialogue management
- Flexible system prompting for behavior control
Frequently Asked Questions
Q: What makes this model unique?
This model uniquely combines instruction-following capabilities with specialized training for creative writing and conversation, using the Metharme prompting format that has proven more effective than traditional approaches.
Q: What are the recommended use cases?
The model is specifically designed for fictional writing and entertainment purposes, including creative writing, roleplay scenarios, and interactive storytelling. It's important to note that it's not intended for factual or safety-critical applications.
