gemma-4-31B-anthology

<div style="background-color: #16171d; color: #e2e8f0; padding: 35px; border-radius: 12px; font-family: system-ui, -apple-system, sans-serif;"> <div align="center"> <h1>📚 Anthology 31B</h1> <p><b>A Gemma-4-31B merge focused on creative storywriting.</b></p> <!-- This is the placeholder for your attached repository image --> <img src="./image.png" alt="Anthology 31B Header Image" width="700px" style="border-radius: 8px; margin: 20px 0; max-width: 100%;"> </div>

---

📌 Overview

Anthology 31B is a creative writing model built via a geometric Model Stock merge of three specialized fine-tunes of gemma-4-31b-it.

Instead of traditional linear weight blending—which can dilute specific behavioral fine-tuning and result in generic text—the Model Stock technique maps the fine-tunes as vectors in a geometric space relative to the base model. This mathematical optimization preserves the distinct literary strengths of each component model rather than averaging them out.

---

🤝 Merged Models

Anthology combines three distinct creative voices to form a singular, cohesive writing instrument:

<table style="width: 100%; border-collapse: collapse; font-size: 0.95rem; margin: 20px 0; background-color: #1e202b; border-radius: 8px; overflow: hidden;"> <thead> <tr style="background-color: #282a36; text-align: left;"> <th style="padding: 15px; border-bottom: 2px solid #3c3f54;">Component Model</th> <th style="padding: 15px; border-bottom: 2px solid #3c3f54;">Primary Contribution</th> </tr> </thead> <tbody> <tr> <td style="padding: 15px; border-bottom: 1px solid #2d3142;"><a href="https://huggingface.co/MRockatansky/Gemma-4-31B-storymaxxed2" style="color: #a78bfa; font-weight: bold; text-decoration: none;">MRockatansky/Gemma-4-31B-storymaxxed2</a></td> <td style="padding: 15px; border-bottom: 1px solid #2d3142;">My own tune, which writes great stories from minimal prompting with a writing style close to stock gemma-4, for better or worse.</td> </tr> <tr> <td style="padding: 15px; border-bottom: 1px solid #2d3142;"><a href="https://huggingface.co/nbeerbower/Gemma4-Gutenberg-31B-Heretic" style="color: #a78bfa; font-weight: bold; text-decoration: none;">nbeerbower/Gemma4-Gutenberg-31B-Heretic</a></td> <td style="padding: 15px; border-bottom: 1px solid #2d3142;">Imbues the output with a classic, novel-like prose structure, and a little dash of antiquated vocabulary from the old public domain fiction it was trained on. </td> </tr> <tr> <td style="padding: 15px; border-bottom: 1px solid #2d3142;"><a href="https://huggingface.co/llmfan46/gemma-4-Ortenzya-The-Creative-Wordsmith-31B-it-uncensored-heretic" style="color: #a78bfa; font-weight: bold; text-decoration: none;">llmfan46/gemma-4-Ortenzya-31B-it-uncensored-heretic</a></td> <td style="padding: 15px; border-bottom: 1px solid #2d3142;">Chosen for its vocabulary choices and writing style.</td> </tr> </tbody> </table>

---

🛠️ Mergekit Configuration

The merge was conducted utilizing zerofata's gemma4-support branch of Mergekit, using the Model Stock merge method.

merge_method: model_stock
base_model: google/gemma-4-31b-it
dtype: bfloat16
normalize: true
models:
  - model: MRockatansky/Gemma-4-31B-storymaxxed2
  - model: nbeerbower/Gemma4-Gutenberg-31B-Heretic
  - model: llmfan46/gemma-4-Ortenzya-The-Creative-Wordsmith-31B-it-uncensored-heretic

---

💜 Special Thanks

• The Beaver AI Discord community: For my introduction to and education in the wonderful world of local open-source AI
• The Tuners: Thanks to nbeerbower and llmfan46 for creating the great models used in this merge
• zerofata: For creating the gemma4-support fork of Mergekit that made this merge possible