FractalSoup-L3-8b : Forbidden Soup

Recommended Settings

Temperature - 1.12-1.22
Min-P - 0.075
Top-K - 50
Repetition Penalty - 1.1
Template : Llama-3-instruct

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 Sao10K/L3-8B-Stheno-v3.2 as a base.

Models Merged

The following models were included in the merge:

• Sao10K/L3-8B-Stheno-v3.2
• SentientAGI/Dobby-Mini-Leashed-Llama-3.1-8B
• Sao10K/L3-8B-Niitama-v1
• SentientAGI/Dobby-Mini-Unhinged-Llama-3.1-8B
• NousResearch/Hermes-3-Llama-3.1-8B

Configuration

The following YAML configuration was used to produce this model:

models:
  - model: SentientAGI/Dobby-Mini-Unhinged-Llama-3.1-8B # generalist, unhinged
    parameters:
      weight: 0.20
      density: 0.5
  - model: NousResearch/Hermes-3-Llama-3.1-8B # generalist, intelligent
    parameters:
      weight: 0.20
      density: 0.5
  - model: Sao10K/L3-8B-Stheno-v3.2 # a good base
    parameters:
      weight: 0.20
      density: 0.5
  - model: SentientAGI/Dobby-Mini-Leashed-Llama-3.1-8B # generalist, leashed
    parameters:
      weight: 0.20
      density: 0.5
  - model: Sao10K/L3-8B-Niitama-v1 # enhanced creativity
    parameters:
      weight: 0.20
      density: 0.5
merge_method: dare_ties
base_model: Sao10K/L3-8B-Stheno-v3.2
parameters:
  normalize: false
out_dtype: bfloat16
chat_template: llama3
tokenizer:
 source: base