Model Overview

WildMBXMarconi-SLERP-7B is a 7 billion parameter language model developed by jsfs11. It is a product of merging two distinct models, BarryFutureman/WildMarcoroni-Variant1-7B and flemmingmiguel/MBX-7B, utilizing the SLERP (Spherical Linear Interpolation) merge method. This approach combines the strengths of its constituent models to enhance overall performance.

Key Capabilities & Performance

This model demonstrates robust performance across a suite of benchmarks, as evaluated on the Open LLM Leaderboard. It achieves an average score of 75.09, with notable results including:

• AI2 Reasoning Challenge (25-Shot): 73.29
• HellaSwag (10-Shot): 88.49
• MMLU (5-Shot): 64.90
• TruthfulQA (0-shot): 68.98
• Winogrande (5-shot): 83.98
• GSM8k (5-shot): 70.89

These scores indicate strong capabilities in reasoning, common sense inference, language understanding, and mathematical problem-solving. The model operates with a context length of 4096 tokens.

Usage

Developers can integrate WildMBXMarconi-SLERP-7B using the Hugging Face transformers library, with provided code examples for text generation. The model is configured for bfloat16 dtype, optimizing for performance and memory efficiency.