ReadyArt/Forgotten-Abomination-24B-V2.2

TEXT GENERATIONConcurrent Unit Cost:2Model Size:24BQuant:FP8Context Size:32kTool Calling:SupportedArchitecture:Transformer0.0K Featherless Exclusive Cold

ReadyArt/Forgotten-Abomination-24B-V2.2 is a 24 billion parameter language model created by ReadyArt, merged using the DARE TIES method. With a 32768-token context length, it combines the strengths of ReadyArt/Forgotten-Safeword-24B-V2.2 and TheDrummer/Cydonia-24B-v2. This model is designed to leverage the combined capabilities of its constituent models for general language tasks.

Loading preview...

Model Overview

ReadyArt/Forgotten-Abomination-24B-V2.2 is a 24 billion parameter language model developed by ReadyArt, utilizing the DARE TIES merge method. This model integrates two distinct pre-trained language models to combine their respective strengths, offering a robust foundation for various natural language processing tasks.

Key Capabilities

  • Merged Architecture: Built upon a base of ReadyArt/Forgotten-Safeword-24B-V2.2, with contributions from TheDrummer/Cydonia-24B-v2, enhancing its overall performance and versatility.
  • Extended Context Window: Features a substantial 32768-token context length, enabling it to process and generate longer, more coherent texts.
  • DARE TIES Method: Employs the DARE TIES (Disentangled and Aligned Representation Ensemble for Task-Independent Scaling) merge method, known for effectively combining models while preserving their individual expertise.

Good For

  • General Language Understanding: Suitable for a broad range of NLP applications requiring strong comprehension and generation capabilities.
  • Complex Text Processing: Its large context window makes it ideal for tasks involving lengthy documents, conversations, or code.
  • Exploratory Development: Provides a solid base for developers looking to experiment with merged model architectures and their emergent properties.

Popular Sampler Settings

Top 3 parameter combinations used by Featherless users for this model. Click a tab to see each config.

temperature
top_p
top_k
frequency_penalty
presence_penalty
repetition_penalty
min_p