Behemoth-123B-v2.1
| Property | Value |
|---|---|
| Parameter Count | 123B |
| Model Type | Mistral-based Language Model |
| Tensor Type | BF16 |
| License | Other |
What is Behemoth-123B-v2.1?
Behemoth-123B-v2.1 is an advanced language model that represents a significant evolution in the BeaverAI model series. It's a sophisticated finetune of the Largestral 2411 architecture, specifically enhanced with system prompt support and creative capabilities. This version (v2.1) is equivalent to the Creative Boost variant of the original series, offering improved performance across all aspects of interaction.
Implementation Details
The model implements a unique frankenformat that combines Metharme with Mistral's system tokens, optimizing its potential for various applications. The implementation requires specific formatting for system prompts and user interactions, ensuring maximum performance.
- Specialized system prompt format with Metharme integration
- Built on Largestral 2411 architecture
- BF16 tensor precision for optimal performance
- Comprehensive system token support
Core Capabilities
- Enhanced creative text generation
- System prompt processing with special token support
- Improved overall performance compared to previous versions
- Flexible integration with various frontend interfaces
Frequently Asked Questions
Q: What makes this model unique?
This model stands out due to its specialized frankenformat implementation and creative boost capabilities, making it particularly effective for creative and general-purpose applications. The integration of Metharme with Mistral's system tokens provides a unique advantage in handling complex prompts.
Q: What are the recommended use cases?
The model excels in creative writing, general text generation, and applications requiring system prompt support. It's particularly well-suited for environments requiring both creative and structured outputs, with special optimization for SillyTavern and similar interfaces.
