Tissint-14B-v1.2-128k-RP
| Property | Value |
|---|---|
| Parameter Count | 14.8B |
| Model Type | Text Generation |
| License | Apache 2.0 |
| Architecture | Transformer-based (Qwen2) |
| Tensor Type | BF16 |
What is Tissint-14B-v1.2-128k-RP?
Tissint-14B is an advanced language model built upon SuperNova-Medius, specifically engineered for enhanced creative text generation with an impressive 128k context window. Version 1.2 represents a significant upgrade with additional training data, processed on NVIDIA A10 hardware.
Implementation Details
The model implements the ChatML format and offers two distinct sampling configurations for different use cases. It features improved stability across various samplers compared to its base model, with notably reduced censorship and better system prompt adherence.
- 128k context window support
- Optimized for multi-turn conversations
- Enhanced creative content generation
- Reduced response length instability
Core Capabilities
- Creative text generation with NSFW support
- Stable performance across different sampling parameters
- Extended context handling
- Improved system prompt adherence
- Multi-turn conversation management
Frequently Asked Questions
Q: What makes this model unique?
The model stands out for its combination of a large context window (128k), enhanced stability across different samplers, and reduced censorship compared to the base model, while maintaining high-quality creative outputs.
Q: What are the recommended use cases?
The model offers two primary configurations: a balanced setting (Temperature 0.8-1.15) for general use, and a creativity setting (Temperature 1.15-1.5) for more imaginative outputs, making it suitable for both standard chatbot applications and creative writing tasks.
