Azazelle/L3-Decent-Lois-Griffin-8B

merge

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

Merge Details

Merge Method

This model was merged using the breadcrumbs_ties merge method using NousResearch/Meta-Llama-3-8B-Instruct as a base.

Models Merged

The following models were included in the merge:

• AwanLLM/Awanllm-Llama-3-8B-Cumulus-v1.0
• DevQuasar/llama3_8b_chat_brainstorm
• NousResearch/Hermes-2-Theta-Llama-3-8B
• chujiezheng/Llama-3-Instruct-8B-SimPO-ExPO
• openchat/openchat-3.6-8b-20240522
• nbeerbower/llama-3-gutenberg-8B
• saishf/Neural-SOVLish-Devil-8B-L3

Configuration

The following YAML configuration was used to produce this model:

models:
  - model: NousResearch/Meta-Llama-3-8B-Instruct
  - model: NousResearch/Hermes-2-Theta-Llama-3-8B # 6/10
    parameters:
      density: 0.4
      weight: 0.15
  - model: openchat/openchat-3.6-8b-20240522 # 5/10
    parameters:
      density: 0.3
      weight: 0.11
  - model: DevQuasar/llama3_8b_chat_brainstorm # 6/10
    parameters:
      density: 0.4
      weight: 0.15
  - model: AwanLLM/Awanllm-Llama-3-8B-Cumulus-v1.0 # 6/10
    parameters:
      density: 0.4
      weight: 0.15
  - model: chujiezheng/Llama-3-Instruct-8B-SimPO-ExPO # 6/10
    parameters:
      density: 0.4
      weight: 0.15
  - model: saishf/Neural-SOVLish-Devil-8B-L3 # 6/10
    parameters:
      density: 0.4
      weight: 0.15
  - model: nbeerbower/llama-3-gutenberg-8B # 6/10
    parameters:
      density: 0.4
      weight: 0.15
merge_method: breadcrumbs_ties
base_model: NousResearch/Meta-Llama-3-8B-Instruct
parameters:
  normalize: false
  rescale: true
  gamma: 0.01
dtype: float16