Cran-May/SCE-2-24B
Cran-May/SCE-2-24B is a 24 billion parameter language model created by Cran-May using the SCE merge method. It is based on huihui-ai/Mistral-Small-24B-Instruct-2501-abliterated and incorporates elements from AlSamCur123/Mistral-Small3-24B-InstructContinuedFine, cognitivecomputations/Dolphin3.0-Mistral-24B, and trashpanda-org/MS-24B-Instruct-Mullein-v0. This model is designed for general instruction-following tasks, leveraging the combined strengths of its merged components.
Loading preview...
Overview
Cran-May/SCE-2-24B is a 24 billion parameter language model developed by Cran-May. It was created using the SCE merge method, a technique detailed in the paper "SCE: A Simple and Effective Method for Merging Pre-trained Language Models". This model is built upon the huihui-ai/Mistral-Small-24B-Instruct-2501-abliterated as its base.
Key Capabilities
- Instruction Following: Designed to respond effectively to a wide range of instructions, benefiting from the instruction-tuned nature of its merged components.
- Merged Architecture: Combines the strengths of multiple pre-trained models, including:
- AlSamCur123/Mistral-Small3-24B-InstructContinuedFine
- cognitivecomputations/Dolphin3.0-Mistral-24B
- trashpanda-org/MS-24B-Instruct-Mullein-v0
- Parameter Efficiency: At 24 billion parameters, it offers a balance between performance and computational requirements, suitable for various applications.
Good for
- General-purpose AI applications: Its broad training across multiple models makes it versatile for diverse tasks.
- Experimentation with merged models: Ideal for developers interested in exploring the capabilities of models created via advanced merging techniques like SCE.
- Instruction-based tasks: Excels in scenarios requiring the model to follow specific commands or prompts.