rityak/NM-Vikhum-Union-12B

merge

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

Merge Details

Merge Method

This model was merged using the DARE TIES merge method using rityak/Vikhr-Nemo-dostoevsky-saiga-12b as a base.

Models Merged

The following models were included in the merge:

• rityak/NM-Vikhr-Magnum-dare-12B
• rityak/Magnum-DareLinearAbliterated-Instruct-DPO
• Aleteian/base-ground-2

Configuration

The following YAML configuration was used to produce this model:

models:
  - model: rityak/Vikhr-Nemo-dostoevsky-saiga-12b
  - model: rityak/NM-Vikhr-Magnum-dare-12B
    parameters:
      weight: .48
      density: [.6, .5, .8]
  - model: rityak/Magnum-DareLinearAbliterated-Instruct-DPO
    parameters:
      weight: .21
      density: [.3, .3, .2]
  - model: Aleteian/base-ground-2
    parameters:
      weight: .21
      density: [.3, .4, .4]
merge_method: dare_ties
tokenizer_source: union
base_model: rityak/Vikhr-Nemo-dostoevsky-saiga-12b
dtype: float16