Veavibork/Anubis-70B-v1.2

Hugging Face
TEXT GENERATIONConcurrent Unit Cost:4Model Size:70BQuant:FP8Context Size:32kTool Calling:SupportedArchitecture:Transformer Featherless Exclusive Warm

Veavibork/Anubis-70B-v1.2 is a 70 billion parameter language model developed by TheDrummer, optimized for creative writing, dynamic storytelling, and entertainment-focused applications. It features a 32768 token context length and is specifically fine-tuned to enhance imaginative capabilities and nuanced language generation, prioritizing user experience in creative use cases over strict alignment or problem-solving intelligence. This model aims to provide a distinct alternative for scenarios requiring expressive and engaging AI interactions.

Loading preview...

Anubis 70B v1.2: A Model for Creative Expression

Anubis 70B v1.2, developed by TheDrummer, is a 70 billion parameter language model specifically fine-tuned for creativity, usability, and entertainment. Unlike foundational models often optimized for non-creative tasks, Anubis 70B v1.2 prioritizes enhancing the user experience for applications requiring imaginative and dynamic language generation. It supports the Llama 3 Chat Template and features a 32768 token context length.

Key Characteristics & Focus Areas

The model's development emphasizes several core qualities:

  • Creativity: Excels in writing style, dynamic storytelling, and imaginative navigation through complex narratives, aiming for pleasant and effective prose.
  • Usability & (Dis)alignment: Designed to avoid common refusal patterns and corporate/religious/political biases, offering flexibility in handling diverse themes and adapting formatting.
  • Intelligence: While not the primary focus, it demonstrates strong adherence to instructions, broad knowledge (fictional and non-fictional), and the ability to handle nuance and logic.

User Feedback Highlights

Reviews indicate that Anubis 70B v1.2 offers a "surprisingly solid" and "much smarter" experience compared to other 70B models, with particular praise for its prose quality, coherent tone, and ability to maintain character speaking mannerisms over long conversations. Users have noted its capacity for creative moments while adhering to character cards and its performance in extended chat scenarios, including pushing context to 32k tokens with coherence.