L3.3-MS-Nevoria-70b
| Property | Value |
|---|---|
| Parameter Count | 70B |
| Base Model | LLaMA 3.3 |
| Author | Steelskull |
| Benchmark Score | UGI Score: 56.75 |
What is L3.3-MS-Nevoria-70b?
L3.3-MS-Nevoria-70b is a sophisticated large language model that combines multiple specialized components to create a powerful storytelling and creative writing system. Built on the LLaMA 3.3 architecture, it integrates EVA-LLAMA-0.1's storytelling capabilities, EURYALE-v2.3's detailed scene descriptions, and Anubis-v1's enhanced prose details, while utilizing Negative_LLAMA to reduce positive bias.
Implementation Details
The model employs a unique "weight twisting" approach through lorablated base model integration, creating distinctive weight interactions similar to the Astoria model series. This unconventional merging technique produces balanced behavioral characteristics and improved response patterns.
- Integrated EVA-LLAMA-0.1 for advanced storytelling
- EURYALE-v2.3 implementation for rich scene descriptions
- Anubis-v1 integration for enhanced prose detail
- Negative_LLAMA components for reduced positive bias
- Nemotron-lorablated base model for unique weight interactions
Core Capabilities
- Exceptional character adherence and dialogue management
- High-quality slow-burn storytelling
- Balanced response generation without positive bias
- Accurate context tracking for complex scenarios
- Support for up to 110,000 tokens without performance degradation
- Strong performance in violence and mature content handling
Frequently Asked Questions
Q: What makes this model unique?
The model's distinctive feature is its balanced approach to storytelling, combining multiple specialized components with a unique weight-twisting technique. It excels in maintaining character consistency and complex plot management while avoiding common issues like positive bias and impersonation problems.
Q: What are the recommended use cases?
The model is particularly well-suited for creative writing, role-playing scenarios, and complex storytelling applications. It performs exceptionally well with detailed character interactions, multiple concurrent plot threads, and scenarios requiring nuanced narrative development.
