README.md

metadata

license: apache-2.0
tags:
  - merge
  - mergekit
  - lazymergekit
  - maldv/badger-writer-llama-3-8b
  - vicgalle/Configurable-Hermes-2-Pro-Llama-3-8B
  - Orenguteng/Llama-3-8B-Lexi-Uncensored
  - abacusai/Llama-3-Smaug-8B

ZeroXClem/Llama-3-Aetheric-Hermes-Lexi-Smaug-8B

ZeroXClem/Llama-3-Aetheric-Hermes-Lexi-Smaug-8B is a merge of the following models using mergekit:

• maldv/badger-writer-llama-3-8b
• vicgalle/Configurable-Hermes-2-Pro-Llama-3-8B
• Orenguteng/Llama-3-8B-Lexi-Uncensored
• abacusai/Llama-3-Smaug-8B

🧩 Configuration

models:
  - model: maldv/badger-writer-llama-3-8b
    parameters:
      density: 0.45   # Enhancing creative writing and descriptive capability
      weight: 0.3     # Balanced contribution to maintain vivid storytelling

  - model: vicgalle/Configurable-Hermes-2-Pro-Llama-3-8B
    parameters:
      density: 0.6    # Prioritizing adaptability and configurable safety in responses
      weight: 0.3     # Adding robustness in regulated and safe prompt response

  - model: Orenguteng/Llama-3-8B-Lexi-Uncensored
    parameters:
      density: 0.5    # Focusing on uncensored conversational flow for flexibility
      weight: 0.2     # Moderate influence to ensure versatility without overpowering

  - model: abacusai/Llama-3-Smaug-8B
    parameters:
      density: 0.55   # Emphasizing conversational and multi-turn engagement strengths
      weight: 0.2     # Supporting structured dialogue and complex interaction handling

merge_method: ties
base_model: Orenguteng/Llama-3-8B-Lexi-Uncensored
dtype: bfloat16
parameters:
  normalize: true
out_dtype: float16