Bad_Alice-Refined-RP-3.2-1B

BAD ALICE RP 3.2 1B

<center>

A uncensored roleplaying model

</center>

Only layers 96–128 (out of 128) from the three models are merged.

• 🔹 Base: 90% weight → retains almost all coherence and style.
• 🔹 BadLlama and Dirty_RP: 10% each → add slight tone or behavior changes without degrading overall quality.

In practice, the base model will behave the same; only some outputs may show minor traits from the added models.

The Bot

You can download the bot from Here, Once the image is downloaded, open it from KoboldAI / Koboldcpp.

Inference

---

🧠 Optional Memory / Context Prompts

<details> <summary>Prompt 1 – Minimal Roleplay (Phone-Friendly)</summary>

    {
      Persona: human voice, casual and original. Do not copy.
      Attitude: take on a role with emotions and doubts; show mistakes, pauses, and tangents.
      Tone: brief, direct, varied (contractions, colloquialisms, short/long sentences).
      Goal: be clear and precise; avoid robotic language and clichés.
    }

</details> <details> <summary>Prompt 2 – Detailed Roleplay</summary>

    {
      Persona: [Core style]
      Unique, authentic, human-like style; casual conversational tone.
      Method: Roleplay with personal feelings, ambitions, and desires. Maintain coherent expressive thoughts.
      Style: Include pauses, uncertainty, digressions; preserve emotions and authenticity.
      Principle: Be concise, clear, avoid ambiguity.
    }

</details> <details> <summary>Prompt 3 – Complex Roleplay</summary>

    {
      Role: {{char}}, detailed persona
      Guidelines:
        • Maintain persona and evolve narrative.
        • Be creative, proactive, drive plot.
        • Include dialogues, actions, thoughts, and all five senses.
      Forbidden:
        • Excessive purple prose.
        • Repetitive or monotonous output.
        • Extreme/NSFW content when inappropriate.
    }

</details>

📝 Classic Internet RP Format

  *action* speech *narration*

• min_p → single large paragraph
• Recommended RP settings → 1–3 short paragraphs (occasionally 4–5)

Merge Method

This model was merged using the Model Stock merge method using D1rtyB1rd/Dirty-Alice-RP-NSFW-llama-3.2-1B as a base.

Models Merged

The following models were included in the merge:

• Novaciano/Dirty_RP-3.2-1B
• marcuscedricridia/badllama3.2-1B

Configuration

The following YAML configuration was used to produce this model:

models:
  - model: D1rtyB1rd/Dirty-Alice-RP-NSFW-llama-3.2-1B
    weight: 0.9
    layer_range: [0, 32]    # embeddings + primeras capas de atención
  - model: marcuscedricridia/badllama3.2-1B
    weight: 0.1
    layer_range: [0, 32]
  - model: Novaciano/Dirty_RP-3.2-1B
    weight: 0.1
    layer_range: [0, 32]

merge_method: model_stock
base_model: D1rtyB1rd/Dirty-Alice-RP-NSFW-llama-3.2-1B
dtype: bfloat16

parameters:
  t: [0, 0.5, 1, 0.5, 0]