SvalTek/Gemma3-ColdBrew-Lorenz-12b-test1

merge

This is a merge of pre-trained language models created using mergekit.

Merge Details

Merge Method

This model was merged using the DELLA merge method using ToastyPigeon/Gemma-3-Starshine-12B-Alt as a base.

Models Merged

The following models were included in the merge:

• SvalTek/Gemma3-ColdBrew-Lorenz
• ToastyPigeon/g3-12b-storyteller-v0.2-textonly

Configuration

The following YAML configuration was used to produce this model:

name: Gemma3-ColdBrew-Lorenz-12b-test1

models:

  - model: SvalTek/Gemma3-ColdBrew-Lorenz
    parameters:
      weight: 0.7
      density: 0.6      # Lower density: more aggressive filtering
      epsilon: 0.18     # High epsilon: only the most relevant changes are kept

  - model: ToastyPigeon/g3-12b-storyteller-v0.2-textonly
    parameters:
      weight: 0.3
      density: 0.75     # Higher density: more skill retained
      epsilon: 0.08     # Lower epsilon: keep broader range of changes

base_model: ToastyPigeon/Gemma-3-Starshine-12B-Alt

merge_method: della
parameters:
  lambda: 1.0
  normalize: true
  int8_mask: true
dtype: bfloat16

tokenizer:
  source: union