TareksLab/Carnelian-SCE-V4-LLaMa-70B

MERGE2

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

Merge Details

Merge Method

This model was merged using the SCE merge method using nbeerbower/Llama-3.1-Nemotron-lorablated-70B as a base.

Models Merged

The following models were included in the merge:

• TareksLab/Larper-V1-LLaMa-70B
• TareksLab/Verbatim-V1-LLaMa-70B
• Neph0s/CoSER-Llama-3.1-70B
• nbeerbower/llama3.1-kartoffeldes-70B
• huihui-ai/DeepSeek-R1-Distill-Llama-70B-abliterated

Configuration

The following YAML configuration was used to produce this model:

models:
  - model: nbeerbower/llama3.1-kartoffeldes-70B
    parameters:
      select_topk: 0.5
  - model: huihui-ai/DeepSeek-R1-Distill-Llama-70B-abliterated
    parameters:
      select_topk: 0.5
  - model: TareksLab/Verbatim-V1-LLaMa-70B
    parameters:
      select_topk: 0.5
  - model: Neph0s/CoSER-Llama-3.1-70B
    parameters:
      select_topk: 0.5
  - model: TareksLab/Larper-V1-LLaMa-70B
    parameters:
      select_topk: 0.5
base_model: nbeerbower/Llama-3.1-Nemotron-lorablated-70B
merge_method: sce
parameters:
dtype: float32
out_dtype: bfloat16
chat_template: llama3
tokenizer:
 source: base
 pad_to_multiple_of: 8