anthracite-org/magnum-v4-27b

Warm
Public
27B
FP8
32768
Oct 20, 2024
License: gemma
Hugging Face
Overview

Model Overview

Magnum-v4-27b is a 27 billion parameter language model developed by Anthracite, built upon the Gemma 27b (chatML'ified) base model. Its primary objective is to emulate the advanced prose quality found in Claude 3 models, specifically Sonnet and Opus, making it suitable for applications requiring sophisticated and nuanced text generation.

Key Capabilities

  • Claude 3 Prose Replication: Fine-tuned to achieve a high standard of prose quality, mirroring the style and depth of Claude 3 models.
  • Extended Context Window: Features a 32768 token context length, allowing for more extensive and coherent conversations or document processing.
  • ChatML Compatibility: Utilizes the ChatML format for prompting, ensuring straightforward integration into existing chat-based applications.
  • Robust Training Data: Trained on a diverse set of datasets, including various Claude-instruct and synthetic generation datasets, to enhance its conversational and creative writing abilities.

Performance Highlights

Evaluations on the Open LLM Leaderboard show an average score of 26.33. Specific metrics include:

  • IFEval (0-Shot): 34.54
  • BBH (3-Shot): 40.96
  • MMLU-PRO (5-shot): 37.51

Use Cases

This model is particularly well-suited for:

  • Conversational AI: Developing chatbots and virtual assistants that require human-like and expressive responses.
  • Creative Writing: Generating stories, role-play scenarios, and other forms of creative content with high prose quality.
  • Advanced Text Generation: Any application where replicating sophisticated writing styles and nuanced language is critical.