Naphula/G4-Runic-Oarfish-26B-A4B-v1.2
Naphula/G4-Runic-Oarfish-26B-A4B-v1.2 is a 26 billion parameter language model based on the Gemma4 architecture, specifically designed for creative roleplay. This model is a merge of Musica, MeroMero, and Animus, utilizing a custom moe_karcher method to combine their strengths. It offers improved prose and reduced refusal behavior compared to previous versions, making it suitable for nuanced conversational and narrative generation with a 32768 token context length.
Loading preview...
G4 Runic Oarfish 26B A4B v1.2 Overview
Naphula/G4-Runic-Oarfish-26B-A4B-v1.2 is a 26 billion parameter model built on the Gemma4 architecture, primarily focused on creative roleplay (RP) applications. This version is a sophisticated merge of three distinct models: Musica, MeroMero, and Darkhn/Gemma-4-26B-A4B-Animus-V14.1-FFT.
Key Capabilities & Features
- Creative Roleplay Optimization: Specifically engineered for generating engaging and nuanced roleplay scenarios and dialogues.
- Advanced Merging Technique: Utilizes a custom
moe_karchermethod, an adaptation of the standard Karcher method for Mixture of Experts (MoE) models, to blend the strengths of its constituent models. - Improved Prose and Reduced Refusals: Version 1.2 demonstrates better writing quality and fewer instances of refusal compared to its predecessors (v1 and v1.1), which exhibited more subtle refusals and shorter replies.
- Gemma4 Architecture: Built upon the Gemma4 base, supporting its underlying capabilities.
- Magnitude Scan Insights: Analysis revealed MeroMero had the strongest influence (highest L2 norm) in the merge direction, followed by Animus and Musica.
When to Use This Model
This model is particularly well-suited for applications requiring:
- Rich, descriptive narrative generation.
- Interactive and creative roleplaying experiences.
- Scenarios where nuanced character interactions are crucial.
While it may exhibit some refusals, these can often be managed through prompt engineering. The model's 32768 token context length further supports complex and extended conversational flows.