Model Overview

arcee-ai/Saul-Instruct-Clown-7b is a 7 billion parameter instruction-tuned language model developed by arcee-ai. It is a merged model, combining the strengths of two distinct base models: CorticalStack/pastiche-crown-clown-7b-dare-dpo and Equall/Saul-Instruct-v1. This merge was performed using the mergekit tool, specifically employing the slerp merge method with defined parameter slices for self-attention and MLP layers.

Key Capabilities & Performance

This model demonstrates competitive performance across a range of benchmarks, as evaluated by the OpenLLM benchmark suite. Its scores include:

• Average: 72.79
• ARC: 68.26
• HellaSwag: 86.28
• MMLU: 63.12
• TruthfulQA: 64.68
• GSM8K: 83.43

These results indicate its proficiency in areas such as common sense reasoning, language understanding, and mathematical problem-solving. The model's 4096-token context length supports handling moderately long inputs and generating coherent responses.

Use Cases

Saul-Instruct-Clown-7b is well-suited for general instruction-following tasks, including:

• Conversational AI: Engaging in dialogue and answering questions.
• Text Generation: Creating various forms of text based on prompts.
• Reasoning Tasks: Assisting with logical deductions and problem-solving where its benchmark performance is strong.