Umbral-Devil-Hermes-Mind-V0.1

merge

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

Merge Details

Merge Method

This model was merged using the TIES merge method using saishf/Neural-SOVLish-Devil-8B-L3 as a base.

Models Merged

The following models were included in the merge:

• Casual-Autopsy/L3-Umbral-Mind-RP-v3.0-8B
• NousResearch/Hermes-3-Llama-3.1-8B

Configuration

The following YAML configuration was used to produce this model:

models:
  - model: Casual-Autopsy/L3-Umbral-Mind-RP-v3.0-8B
    parameters:
      density: 0.5
      weight: 0.4
      enhanced_attention: true
      abstract_attention: true
      deep_cognitive_focus: true
      dynamic_attention_allocation: true
      significance_threshold: 0.85
      feedback_consciousness: true
      non_linear_resonance: true
      attention_heads:
        - layer_range: [0, 8]
          value: 32
          resonance_amplification: true
        - layer_range: [8, 16]
          value: 28
          resonance_amplification: true
        - layer_range: [16, 24]
          value: 20
          adaptive_significance: true
        - layer_range: [24, 32]
          value: 16
          significance_suppression: true

  - model: NousResearch/Hermes-3-Llama-3.1-8B
    parameters:
      density: 0.4
      weight: 0.5
      long_term_attention: true
      task_specialization: true
      semantic_linking: true
      attention_resonance: true
      focus_regulation: true
      feedback_consciousness: true
      adaptive_resonance_control: true
      attention_heads:
        - layer_range: [0, 8]
          value: 32
          resonance_amplification: true
        - layer_range: [8, 16]
          value: 24
          resonance_amplification: true
        - layer_range: [16, 24]
          value: 16
          adaptive_significance: true
        - layer_range: [24, 32]
          value: 12
          significance_suppression: true

  - model: saishf/Neural-SOVLish-Devil-8B-L3
    parameters:
      density: 0.3
      weight: 0.5
      enhanced_attention: true
      abstract_attention: true
      deep_cognitive_focus: true
      dynamic_attention_allocation: true
      significance_threshold: 0.8
      feedback_consciousness: true
      non_linear_resonance: true
      attention_heads:
        - layer_range: [0, 8]
          value: 32
          resonance_amplification: true
        - layer_range: [8, 16]
          value: 28
          resonance_amplification: true
        - layer_range: [16, 24]
          value: 20
          adaptive_significance: true
        - layer_range: [24, 32]
          value: 16
          significance_suppression: true

merge_method: ties
base_model: saishf/Neural-SOVLish-Devil-8B-L3
parameters:
  normalize: false
  int8_mask: true
  significance: 0.85
  optimal_attention_threshold: 0.9
dtype: bfloat16