GRRNMAKE/Magnus
GRRNMAKE/Magnus is a 7 billion parameter language model fine-tuned from Mistral-7B-Instruct-v0.3 by GRRNMAKE. It is specifically optimized for rigorous tool calling, precise formatting adherence, and extreme conciseness. This model excels at generating accurate and compact responses, making it suitable for applications requiring strict output formats and efficient tool integration.
Loading preview...
GRRNMAKE/Magnus: Optimized for Tool Calling and Conciseness
Magnus is a specialized 7 billion parameter language model developed by GRRNMAKE, built upon the mistralai/Mistral-7B-Instruct-v0.3 base architecture. It has undergone fine-tuning using 4-bit QLoRA with the GRRNMAKER/axim-alignment-data (Axim Corrective Benchmark) dataset, focusing on enhancing its capabilities in tool calling, formatting adherence, and response conciseness.
Key Capabilities & Performance
- Exceptional Tool Calling: Optimized for generating precise and functional tool calls.
- High Formatting Adherence: Achieves verified formatting fidelity, crucial for structured outputs.
- Extreme Conciseness: Demonstrates a 98.2% concise response accuracy and a 94.00 conciseness score on specific benchmarks.
- Low Hallucination Rate: Designed to produce reliable and factual outputs.
- Instruction Adherence: Exhibits high adherence to given instructions, ensuring predictable behavior.
Ideal Use Cases
- Automated Workflows: Excellent for systems requiring strict JSON or API call formats.
- Agentic Applications: Suited for AI agents that need to interact with external tools reliably.
- Structured Data Generation: Effective in scenarios where output conciseness and format are paramount.
- Efficient API Integration: Simplifies integration with APIs by consistently generating correct and minimal payloads.