Model Overview

anthracite-org/magnum-v2-12b is a 12 billion parameter language model developed by Anthracite, built upon the Mistral-Nemo-Base-2407 architecture. This model is the fourth iteration in a series focused on emulating the high prose quality of Claude 3 models, specifically Sonnet and Opus.

Key Capabilities

• Advanced Prose Generation: Fine-tuned to replicate the sophisticated writing style and nuance found in Claude 3 models.
• Instruction Following: Instruct-tuned using ChatML formatting for clear and effective interaction.
• Robust Training: Fine-tuned for 2 epochs using 8 NVIDIA H100 Tensor Core GPUs, leveraging datasets like Stheno, Opus_Instruct_25k, Opus_WritingStruct, and a subset of Sonnet3.5-SlimOrcaDedupCleaned.

Use Cases

• Sophisticated Content Creation: Ideal for generating high-quality articles, stories, or conversational responses that require advanced linguistic capabilities.
• Conversational AI: Suitable for chatbots and virtual assistants where nuanced and human-like dialogue is crucial.
• Prose Replication: Can be used in applications aiming to mimic specific high-quality writing styles.

Performance Metrics

Evaluations on the Open LLM Leaderboard show an average score of 18.68, with specific results including 37.62 on IFEval (0-Shot) and 28.79 on BBH (3-Shot).