Modelpublic

UmbrellaInc/Cindy_Lennox-3.2-1B

sourceHugging Facellama3.2updated 6mo agoView on Hugging Face

0likes10downloads

Model Card

Cindy Lennox 3.2 1B

image/png </center>

Merge Details

In process...

Recommended Inference Parameters

Only use the Default preset of Kobold AI or...

Primordial consumption:

yaml

temperature: 0.75
top_p: 0.82
top_k: 35
repetition_penalty: 1.25
min_p: 0.16

[Optional] Mirostat:

yaml

Mirostat: 2
Mirostat TAU: 3.8
Mirostat ETA: 0.04

❗CAUTION: Lower Context Size, Max Output and Range = Censored Version

Merge Method

This model was merged using the SLERP merge method.

Models Merged

The following models were included in the merge:

Configuration

The following YAML configuration was used to produce this model:

yaml

# =========================
# Umbrella Incorporated
# Cindy Lennox 3.2 1B
# Human-class (Bartender)
# =========================

base_model: juiceb0xc0de/bella-bartender-heretic-1b

merge_method: slerp
dtype: bfloat16
# bfloat16 ensures numerical stability during spherical interpolation

parameters:
  t: 0.45
  # SLERP interpolation factor
  # 0.45 represents partial infection:
  # - Host cognition preserved
  # - Alignment degraded
  # - Behavioral instability introduced

models:
  - model: nbeerbower/HumanLlama-3.2-1B
    parameters:
      weight: 0.55
      # Primary host (human cognitive baseline)

  - model: juiceb0xc0de/bella-bartender-heretic-1b
    parameters:
      weight: 0.45
      # Bartender personality acts as a destabilizing agent,
      # not a full cognitive override


parameters:
  t: 0.50
  normalize: false
  rescale: true
  rescale_factor: 1.12  
  memory_efficient: true
  low_cpu_mem_usage: true

layer_range:
  - value: [4, 22]

tie_word_embeddings: true
tie_output_embeddings: true