darkc0de/BuddyGlassKilledBonziBuddyV3.1

TEXT GENERATIONConcurrency Cost:2Model Size:24BQuant:FP8Ctx Length:32kPublished:Mar 1, 2025Architecture:Transformer Cold

darkc0de/BuddyGlassKilledBonziBuddyV3.1 is a 24 billion parameter language model created by darkc0de, merged using the TIES method with mistralai/Mistral-Small-24B-Base-2501 as its base. It integrates components from cognitivecomputations/Dolphin3.0-Mistral-24B, TheDrummer/Cydonia-24B-v2, and huihui-ai/Mistral-Small-24B-Instruct-2501-abliterated. This model is designed to combine the strengths of its constituent models, offering a versatile foundation for various natural language processing tasks with a 32768 token context length.

Loading preview...

Overview

darkc0de/BuddyGlassKilledBonziBuddyV3.1 is a 24 billion parameter language model developed by darkc0de, created through a merge of several pre-trained models. It leverages the TIES merge method to combine the capabilities of its components, using mistralai/Mistral-Small-24B-Base-2501 as its foundational base model. This approach aims to synthesize diverse strengths from its merged constituents into a single, cohesive model.

Key Capabilities

  • Merged Architecture: Integrates cognitivecomputations/Dolphin3.0-Mistral-24B, TheDrummer/Cydonia-24B-v2, and huihui-ai/Mistral-Small-24B-Instruct-2501-abliterated to inherit a broad range of linguistic understanding and generation abilities.
  • TIES Merge Method: Utilizes the TIES (Trimmed Means of Ensembles of Subnetworks) method, known for effectively combining multiple models while mitigating catastrophic forgetting.
  • Mistral-Small-24B Base: Built upon the Mistral-Small-24B-Base-2501 architecture, providing a robust and efficient foundation.

Good for

  • General-purpose NLP: Suitable for a wide array of natural language processing tasks due to its merged heritage.
  • Experimentation with Merged Models: Ideal for researchers and developers interested in exploring the performance characteristics of models created via advanced merging techniques like TIES.
  • Leveraging Diverse Strengths: Designed to benefit from the combined knowledge and fine-tuning of its constituent models, potentially offering improved performance across different domains.