grimjim/Magrathic-12B

TEXT GENERATIONConcurrency Cost:1Model Size:12BQuant:FP8Ctx Length:32kPublished:Jul 1, 2025License:cc-by-nc-4.0Architecture:Transformer0.0K Open Weights Cold

Magrathic-12B is a 12 billion parameter language model developed by grimjim, created through a Task Arithmetic merge of several base models, including grimjim/mistralai-Mistral-Nemo-Base-2407 and grimjim/Magnolia-Mell-v1-12B. It incorporates cgato/Nemo-12b-Humanize-SFT-v0.2.5-KTO at a low weight to enhance human-like text generation. This model is designed for general text generation tasks, with a particular focus on producing more humanized and natural language outputs, and supports a 32768 token context length.

Loading preview...

Magrathic-12B Overview

Magrathic-12B is a 12 billion parameter language model developed by grimjim, built upon the Mistral-Nemo-Base-2407 architecture. It was created using the Task Arithmetic merge method, combining several pre-trained models to achieve its specific characteristics. The primary goal of this merge was to enhance the model's ability to generate text that is more human-like and natural in its phrasing.

Key Capabilities

  • Humanized Text Generation: A small weight from cgato/Nemo-12b-Humanize-SFT-v0.2.5-KTO was specifically included to perturb outputs towards more human-like text.
  • Merged Architecture: Integrates capabilities from grimjim/Magnolia-Mell-v1-12B and grimjim/mistralai-Mistral-Nemo-Base-2407.
  • Context Length: Supports a substantial context window of 32768 tokens, allowing for processing longer inputs and generating coherent extended responses.
  • Prompt Format Compatibility: Utilizes the prompt format established by the original Nemo Instruct 2407 12B model.

When to Use Magrathic-12B

This model is particularly well-suited for applications requiring:

  • General Text Generation: Creating diverse and coherent text across various topics.
  • Enhanced Naturalness: Scenarios where the output needs to sound more conversational and less robotic.
  • Creative Writing: Generating stories, dialogues, or other creative content where human-like expression is valued.
  • Long-form Content: Leveraging its large context window for tasks involving extensive input or output.

Finetunes
2 models