Sorihon/Chaotic-Order-24B-V2
Sorihon/Chaotic-Order-24B-V2 is a 24 billion parameter language model created by Sorihon, merged using the DARE TIES method. This model combines Sorihon/Magistry-Painted-Cydonia-24B and DarkArtsForge/Asmodeus-24B-v3, building upon Sorihon/Chaotic-Order-24B-V1 as its base. With a context length of 32768 tokens, it leverages the strengths of its constituent models to offer a versatile foundation for various language generation tasks. Its unique merging approach aims to synthesize diverse capabilities from its components.
Loading preview...
Chaotic-Order-24B-V2: A Merged Language Model
Chaotic-Order-24B-V2 is a 24 billion parameter language model developed by Sorihon. It is the result of a sophisticated merge operation, building upon the previous version, Sorihon/Chaotic-Order-24B-V1, as its base. This model integrates capabilities from multiple distinct pre-trained models to create a new, potentially more versatile, language understanding and generation system.
Merge Details and Architecture
The model was constructed using the DARE TIES merge method, a technique designed to combine the strengths of different models effectively. The merging process involved specific configurations for density and weight for each contributing model, ensuring a balanced integration. The primary models merged into Chaotic-Order-24B-V2 include:
- Sorihon/Magistry-Painted-Cydonia-24B
- DarkArtsForge/Asmodeus-24B-v3
These models were combined with Sorihon/Chaotic-Order-24B-V1 serving as the foundational base. The merge was performed using mergekit, a tool for creating new models from existing ones. The model operates with bfloat16 precision.
Potential Use Cases
Given its merged nature, Chaotic-Order-24B-V2 is likely suitable for a broad range of general-purpose language tasks, benefiting from the combined knowledge and capabilities of its constituent models. Developers looking for a model that synthesizes different strengths from established 24B models may find this merge particularly useful.