models:
  - model: Sao10K/Llama-3.1-8B-Stheno-v3.4+grimjim/Llama-3-Instruct-abliteration-LoRA-8B
    parameters:
      weight: 1.0
merge_method: linear
dtype: float32