Model Overview

saishf/West-Maid-7B is a 7 billion parameter language model developed by saishf. It was created using the SLERP merge method from MergeKit, combining the strengths of two distinct base models: senseable/WestLake-7B-v2 and NeverSleep/Noromaid-7B-0.4-DPO.

Key Capabilities & Performance

This merged model demonstrates a balanced performance across various benchmarks, as evaluated on the Open LLM Leaderboard. It achieved an average score of 69.09, with notable results including:

• AI2 Reasoning Challenge (25-Shot): 67.24
• HellaSwag (10-Shot): 86.44
• MMLU (5-Shot): 64.85
• TruthfulQA (0-shot): 51.00
• Winogrande (5-shot): 82.72
• GSM8k (5-shot): 62.32

These scores indicate proficiency in reasoning, common sense, factual recall, and mathematical problem-solving. The model's 4096-token context length supports processing moderately long inputs.

When to Use This Model

West-Maid-7B is suitable for applications requiring a versatile 7B parameter model that benefits from the combined characteristics of its constituent models. Its balanced performance makes it a strong candidate for:

• General text generation and understanding tasks.
• Reasoning and question-answering scenarios.
• Applications where a blend of capabilities from WestLake-7B-v2 and Noromaid-7B-0.4-DPO is desired.