metadata

base_model:
  - InferenceIllusionist/SorcererLM-22B
  - TheDrummer/Cydonia-22B-v1.3
  - unsloth/Mistral-Small-Instruct-2409
  - Saxo/Linkbricks-Horizon-AI-Japanese-Superb-V1-22B
  - anthracite-org/magnum-v4-22b
  - ArliAI/Mistral-Small-22B-ArliAI-RPMax-v1.1
  - spow12/ChatWaifu_v2.0_22B
  - unsloth/Mistral-Small-Instruct-2409
  - rAIfle/Acolyte-LORA
  - byroneverson/Mistral-Small-Instruct-2409-abliterated
  - Gryphe/Pantheon-RP-Pure-1.6.2-22b-Small
  - allura-org/MS-Meadowlark-22B
  - crestf411/MS-sunfall-v0.7.0
  - TheDrummer/Cydonia-22B-v1.1
  - TheDrummer/Cydonia-22B-v1.2
  - unsloth/Mistral-Small-Instruct-2409
  - Kaoeiri/Moingooistrial-22B-V1-Lora
library_name: transformers
tags:
  - mergekit
  - merge

merge

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 unsloth/Mistral-Small-Instruct-2409 as a base.

Models Merged

The following models were included in the merge:

Configuration

The following YAML configuration was used to produce this model:

models:
  - model: anthracite-org/magnum-v4-22b
    parameters:
      weight: 1.0         # Primary model for human-like writing
      density: 0.88       # Strong foundation with room for blending
  - model: TheDrummer/Cydonia-22B-v1.3
    parameters:
      weight: 0.27        # Balanced for creative flair
      density: 0.71       # Subtle creativity with strong coherence
  - model: TheDrummer/Cydonia-22B-v1.2
    parameters:
      weight: 0.17        # Light creativity for nuanced diversity
      density: 0.68       # Maintains alignment with overarching structure
  - model: TheDrummer/Cydonia-22B-v1.1
    parameters:
      weight: 0.2         # Adds depth to accurate and specific nuances
      density: 0.69       # Smoothly integrates details without overwhelming
  - model: Gryphe/Pantheon-RP-Pure-1.6.2-22b-Small
    parameters:
      weight: 0.3         # Refined for deeper storytelling and RP focus
      density: 0.78       # Supports narrative without clashing
  - model: allura-org/MS-Meadowlark-22B
    parameters:
      weight: 0.29        # Balances creativity with structured fluency
      density: 0.72       # Enhances clarity and descriptive depth
  - model: spow12/ChatWaifu_v2.0_22B
    parameters:
      weight: 0.27        # Maintains anime-style RP and conversational tone
      density: 0.7        # Intact for balanced integration
  - model: Saxo/Linkbricks-Horizon-AI-Japanese-Superb-V1-22B
    parameters:
      weight: 0.19        # Specialized for Japanese contexts
      density: 0.58       # Ensures contextual accuracy without overlap
  - model: crestf411/MS-sunfall-v0.7.0
    parameters:
      weight: 0.26        # Enhanced for impactful dramatic storytelling
      density: 0.74       # Balances spicy narratives with other models
  - model: unsloth/Mistral-Small-Instruct-2409+rAIfle/Acolyte-LORA
    parameters:
      weight: 0.25        # Balanced for varied structured content
      density: 0.71       # Ensures seamless alignment with base
  - model: InferenceIllusionist/SorcererLM-22B
    parameters:
      weight: 0.22        # Stylized refinement for cohesive outputs
      density: 0.73       # Keeps stylistic diversity in balance
  - model: unsloth/Mistral-Small-Instruct-2409+Kaoeiri/Moingooistrial-22B-V1-Lora
    parameters:
      weight: 0.24        # Mythical storytelling integration
      density: 0.71       # Balanced for smooth interaction
  - model: ArliAI/Mistral-Small-22B-ArliAI-RPMax-v1.1
    parameters:
      weight: 0.1         # Light touch to avoid excessive RP influence
      density: 0.64       # Fine-tuned for roleplay-specific elements
  - model: byroneverson/Mistral-Small-Instruct-2409-abliterated
    parameters:
      weight: 0.16        # Adds raw and unfiltered context nuance
      density: 0.69       # Supports diverse content without overpowering

merge_method: dare_ties  # Best for diverse and complex model blending
base_model: unsloth/Mistral-Small-Instruct-2409
parameters:
  density: 0.85          # Overall density ensures logical and creative balance
  epsilon: 0.08          # Reduced for smoother model interpolation
  lambda: 1.23           # Balanced scaling for crisp and coherent outputs
dtype: bfloat16