mmoza32/ShadowLM-Final-Core

TEXT GENERATIONConcurrency Cost:1Model Size:8BQuant:FP8Ctx Length:8kPublished:Apr 3, 2026Architecture:Transformer Cold

ShadowLM-Final-Core by mmoza32 is an 8 billion parameter language model with an 8192-token context length, created by merging several Llama-3.1-8B-Instruct variants using the Linear merge method. This model leverages the strengths of NousResearch/Hermes-3-Llama-3.1-8B, unsloth/Meta-Llama-3.1-8B-Instruct, and mlabonne/Meta-Llama-3.1-8B-Instruct-Abliterated. It is designed to combine the capabilities of its constituent models, offering a versatile foundation for various instruction-following tasks.

Loading preview...

ShadowLM-Final-Core Overview

ShadowLM-Final-Core is an 8 billion parameter language model developed by mmoza32, built upon the Llama-3.1-8B architecture. It was created using the mergekit tool, specifically employing the Linear merge method to combine multiple pre-trained models.

Key Capabilities

This model integrates the characteristics of three distinct Llama-3.1-8B-Instruct variants:

  • NousResearch/Hermes-3-Llama-3.1-8B: Served as the base model for the merge, contributing 40% of the weight.
  • unsloth/Meta-Llama-3.1-8B-Instruct: Included with a 30% weight, enhancing instruction-following capabilities.
  • mlabonne/Meta-Llama-3.1-8B-Instruct-Abliterated: Also contributed 30% of the weight, likely bringing specialized instruction-tuned performance.

By combining these models, ShadowLM-Final-Core aims to offer a robust and versatile solution for general-purpose instruction-based tasks, benefiting from the collective strengths of its components. Its 8192-token context length supports processing longer inputs and generating more extensive responses.

When to Use This Model

ShadowLM-Final-Core is suitable for developers seeking an 8B parameter model that:

  • Requires strong instruction-following abilities.
  • Benefits from a blend of different Llama-3.1-8B-Instruct fine-tunes.
  • Needs a model with a standard 8K context window for various applications.