CoolFace
Modelpublic

Ateron/Gemma-4-Dark-Thoughts-V2-31B

sourceHugging Faceapache-2.0updated 2mo agoView on Hugging Face
23likes381downloads
Model Card

<div style="background: linear-gradient(120deg, #061116 0%, #061610 45%, #05242e 100%); color: #f0f2f5; padding: 30px; border-radius: 12px; border: 4px solid #040c1a; box-shadow: 0 8px 25px rgba(0,0,0,0.9);">

<h1 align="center" style="color: #1baaa3; text-shadow: 2px 2px 5px #000000;">🚬 Dark-Thoughts V2</h1>

<p align="center"> <img src= "https://cdn-uploads.huggingface.co/production/uploads/6390f888de25f9eda571b0f2/A3b0Qc08UUDJZ3igu3HJn.png" width="100%" alt="Dark-Thoughts AI Persona" style="border-radius: 8px; border: 2px solid #061610; box-shadow: 0 4px 15px rgba(0,0,0,0.7);"> </p>

<p align="center" style="font-style: italic; color: #1baaa3; margin-top: 15px; margin-bottom: 25px; font-size: 1.15em; text-shadow: 1px 1px 3px rgba(0,0,0,0.7);"> "Shady corners of mind" </p>

<h2 style="color: #1baaa3; border-bottom: 1px solid #13455c; padding-bottom: 8px; margin-top: 30px;">🎥 Overview</h2>

So, despite creativity the first version had issues with tool calling and brains in long context role-play. I fixed this in V2 and changed recipe, putting more experiments on focus and filters.

<h3 style="color: #1baaa3; margin-top: 15px;">💿 Mergekit Configuration</h3>

Below is the exact mergekit_config.yml recipe used to synthesize this model:

Phase 1: Spark

Ignition of creativity with MeroMero V2 and Dark-Scarlett V2

<details>

<summary>dare_ties Recipe 1</summary>

merge_method: dare_ties
base_model: F:\AI\Merge\Gemma-4-it
tokenizer_source: base
dtype: bfloat16
parameters:
  lambda: 1.0
models:
  - model: F:\AI\Merge\Gemma-4-MeroMero-V2
    parameters:
      density: 0.50
      weight:
        - filter: "model.layers.4."
          value: 0.60
        - filter: "model.layers.9."
          value: 0.60
        - filter: "model.layers.14."
          value: 0.60
        - filter: "model.layers.19."
          value: 0.60
        - filter: "model.layers.24."
          value: 0.60
        - filter: "model.layers.29."
          value: 0.60
        - filter: "model.layers.34."
          value: 0.60
        - filter: "model.layers.39."
          value: 0.50
        - filter: "model.layers.44."
          value: 0.50
        - filter: "model.layers.49."
          value: 0.40
        - filter: "model.layers.54."
          value: 0.40
        - filter: "model.layers.59."
          value: 0.40
        - value: 0.50

  - model: F:\AI\Merge\Gemma-4-Dark-Scarlett-V2
    parameters:
      density: 0.50
      weight:
        - filter: "model.layers.4."
          value: 0.40
        - filter: "model.layers.9."
          value: 0.40
        - filter: "model.layers.14."
          value: 0.40
        - filter: "model.layers.19."
          value: 0.40
        - filter: "model.layers.24."
          value: 0.40
        - filter: "model.layers.29."
          value: 0.40
        - filter: "model.layers.34."
          value: 0.40
        - filter: "model.layers.39."
          value: 0.50
        - filter: "model.layers.44."
          value: 0.50
        - filter: "model.layers.49."
          value: 0.60
        - filter: "model.layers.54."
          value: 0.60
        - filter: "model.layers.59."
          value: 0.60
        - value: 0.50

</details>

Phase 2: Form

Polisher of thoughts from Scotoma v2

<details>

<summary>dare_ties Recipe 2</summary>

merge_method: dare_ties
base_model: F:\AI\Merge\Gemma-4-it
tokenizer_source: base
dtype: bfloat16
parameters:
  lambda: 1.0
models:
  - model: F:\AI\Merge\Gemma-4-Scotoma-V2
    parameters:
      density: 0.60
      weight:
      - {filter: mlp, value: 0.2}
      - {filter: self_attn, value: 0.2}
      - {filter: q_proj, value: 0.5}
      - {filter: k_proj, value: 0.5}
      - {filter: v_proj, value: 0.6}
      - {filter: o_proj, value: 0.8}
      - value: [0.60, 0.60, 0.70, 0.80, 0.80]
  - model: F:\AI\Merge\Dark-Mero-V3
    parameters:
      density: 0.40
      weight:
      - {filter: mlp, value: 0.8}
      - {filter: self_attn, value: 0.8}
      - {filter: q_proj, value: 0.5}
      - {filter: k_proj, value: 0.5}
      - {filter: v_proj, value: 0.4}
      - {filter: o_proj, value: 0.2}
      - value: [0.40, 0.40, 0.30, 0.20, 0.20]

</details>

<h2 style="color: #1baaa3; border-bottom: 1px solid #13455c; padding-bottom: 8px; margin-top: 30px;">🤝 Special Thanks</h2>

<ul style="color: #e2e4e9; line-height: 1.6;"> <li><strong>Google DeepMind:</strong> For providing the base model.</li> <li><strong>The ReadyArt:</strong> Team of good people and creators of Scotoma and Dark Scarlett.</li> <li><strong>To Zerofata:</strong> For keeping merging possible and for wonderful MeroMero 2.</li> <li><strong>To the community:</strong> For giving feedbacks and positivity.</li> <li><strong>To Nimbz:</strong> <a href="https://huggingface.co/Nimbz" style="color: #1baaa3; text-decoration: none;">This cat</a> for being around in tough times .</li> </ul>

</div>