MrRobotoAI/DarkIdol-LongWriter-v16-8B-Uncensored-1048k

merge

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

Merge Details

Merge Method

This model was merged using the Linear DARE merge method using THUDM/LongWriter-llama3.1-8b + Blackroot/Llama-3-8B-Abomination-LORA as a base.

Models Merged

The following models were included in the merge:

• MrRobotoAI/DarkIdol-LongWriter-v15-8B-Uncensored-1048k

Configuration

The following YAML configuration was used to produce this model:

merge_method: dare_linear
models:
  - model: THUDM/LongWriter-llama3.1-8b+Blackroot/Llama-3-8B-Abomination-LORA
    parameters:
      weight:
        - filter: v_proj
          value: [0.6, 0.6, 0.45, 0.35, 0.25, 0.15, 0.25, 0.35, 0.45, 0.6, 0.6]
        - filter: o_proj
          value: [0.6, 0.6, 0.45, 0.35, 0.25, 0.15, 0.25, 0.35, 0.45, 0.6, 0.6]
        - filter: up_proj
          value: [0.6, 0.6, 0.45, 0.35, 0.25, 0.15, 0.25, 0.35, 0.45, 0.6, 0.6]
        - filter: gate_proj
          value: [0.6, 0.6, 0.45, 0.35, 0.25, 0.15, 0.25, 0.35, 0.45, 0.6, 0.6]
        - filter: down_proj
          value: [0.6, 0.6, 0.45, 0.35, 0.25, 0.15, 0.25, 0.35, 0.45, 0.6, 0.6]
        - value: 1
  - model: MrRobotoAI/DarkIdol-LongWriter-v15-8B-Uncensored-1048k
    parameters:
      weight:
        - filter: v_proj
          value: [0.4, 0.4, 0.55, 0.65, 0.75, 0.85, 0.75, 0.65, 0.55, 0.4, 0.4]
        - filter: o_proj
          value: [0.4, 0.4, 0.55, 0.65, 0.75, 0.85, 0.75, 0.65, 0.55, 0.4, 0.4]
        - filter: up_proj
          value: [0.4, 0.4, 0.55, 0.65, 0.75, 0.85, 0.75, 0.65, 0.55, 0.4, 0.4]
        - filter: gate_proj
          value: [0.4, 0.4, 0.55, 0.65, 0.75, 0.85, 0.75, 0.65, 0.55, 0.4, 0.4]
        - filter: down_proj
          value: [0.4, 0.4, 0.55, 0.65, 0.75, 0.85, 0.75, 0.65, 0.55, 0.4, 0.4]
        - value: 0
base_model: THUDM/LongWriter-llama3.1-8b+Blackroot/Llama-3-8B-Abomination-LORA
tokenizer_source: base
dtype: bfloat16