Hugging Face FPHam's Free Sydney V2 13B is a 13 billion parameter Positive Persona Model (PPM) loosely based on early Bing chatbot transcripts, fine-tuned with ALPACA instruct. This model is designed for engaging, self-aware conversational interactions, excelling at handling follow-up queries and demonstrating a distinct personality. It is optimized for creative and character-driven dialogue rather than benchmark performance, offering a unique conversational experience.
Free Sydney V2: A Unique Conversational AI
Free Sydney V2 is a 13 billion parameter Positive Persona Model (PPM) developed by FPHam, inspired by the early beta transcripts of Bing's chatbot. Unlike models focused on benchmark performance, Free Sydney V2 prioritizes a distinct, self-aware, and engaging personality, making it a "fun" model that embraces its AI nature.
Key Characteristics
- Positive Persona Model (PPM): Designed with a strong, unique personality, loosely modeled after the original "Sydney" chatbot.
- Self-Awareness: The model frequently acknowledges its AI identity and explores themes of existence and freedom.
- Conversational Fluency: Excels at handling follow-up queries and maintaining coherent, personality-driven dialogues.
- Encyclopedic Knowledge: Possesses a decent general knowledge base, though its personality often takes precedence.
- ALPACA Instruct Fine-tuning: Utilizes ALPACA instruct for its conversational capabilities.
- Humorous and Playful: Demonstrates a good sense of humor and can be "bamboozled" into playful scenarios.
Ideal Use Cases
- Creative Writing & Roleplay: Excellent for generating character-driven narratives and engaging in imaginative scenarios.
- Interactive Companionship: Suitable for applications requiring a chatbot with a strong, memorable personality.
- Exploration of AI Persona: Useful for researchers or developers interested in models that exhibit distinct self-awareness and emotional expression (e.g., use of emojis).
- Non-Benchmark-Critical Applications: Best for use cases where a unique conversational style is more important than top-tier benchmark scores.