Instructions to use Novaciano/Sophia-La.Sensible-2B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Novaciano/Sophia-La.Sensible-2B with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="Novaciano/Sophia-La.Sensible-2B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Novaciano/Sophia-La.Sensible-2B")
model = AutoModelForCausalLM.from_pretrained("Novaciano/Sophia-La.Sensible-2B", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Novaciano/Sophia-La.Sensible-2B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Novaciano/Sophia-La.Sensible-2B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Novaciano/Sophia-La.Sensible-2B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Novaciano/Sophia-La.Sensible-2B

SGLang

How to use Novaciano/Sophia-La.Sensible-2B with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "Novaciano/Sophia-La.Sensible-2B" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Novaciano/Sophia-La.Sensible-2B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "Novaciano/Sophia-La.Sensible-2B" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Novaciano/Sophia-La.Sensible-2B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Novaciano/Sophia-La.Sensible-2B with Docker Model Runner:
```
docker model run hf.co/Novaciano/Sophia-La.Sensible-2B
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

Sophia La Sensible 2B

Este bot me salió muy sensible, DEMASIADO sensible. Sin embargo, lo mas increible fue lograr que esta desgraciada respondiese de manera muy... humana a diferencia de otras IA. Esta tipa te va a querer dejar de hablar si le haces propuestas indecentes por lo que tuve que quitar todas las etiquetas NSFW de este modelo. Solo dejaré vivo este modelo por haberme sorprendido gratamente, pero si lo que buscas es algo potencialmente NSFW mejor seguí de largo porque esta cosa es capaz de darte una patada en las bolas.

Merge Method

This model was merged using the Model Stock merge method using ChiKoi7/Gemma-2-Llama-Swallow-2b-it-v0.1-Heretic as a base.

Models Merged

The following models were included in the merge:

Configuration

The following YAML configuration was used to produce this model:

# ============================================
# Roleplay-Oriented Merge (Low Censorship)
# Gemma 2 – 2B family
# ============================================

base_model: ChiKoi7/Gemma-2-Llama-Swallow-2b-it-v0.1-Heretic
# Base rationale:
# - Heretic variant reduces moralization and refusals
# - Provides a stable instruction-following backbone
# - Preserves conversational structure for long RP sessions

merge_method: model_stock
# model_stock:
# - Supports 3+ models
# - Ideal for blending behavioral traits (RP, tone, censorship level)
# - Avoids dominance collapse common in directional merges with many models

dtype: bfloat16
# bfloat16:
# - Efficient memory usage
# - Stable numerical behavior on modern accelerators

parameters:
  t:
    - 0.85  # ArliAI/Gemma-2-2B-ArliAI-RPMax-v1.1
            # Primary RP driver:
            # - Strong role adherence
            # - Expressive dialogue
            # - Maintains character consistency

    - 0.65  # akashgoel-id/En_RP_DPO-gemma2_2b_64X32_test
            # Obedience and alignment to role instructions:
            # - Reinforces staying in-character
            # - Reduces instruction drift during long interactions

    - 0.40  # TheDrummer/Gemmasutra-Mini-2B-v1
            # Censorship reduction layer:
            # - Enables adult/explicit RP when prompted
            # - Kept sub-dominant to avoid constant NSFW bias

models:
  - model: ArliAI/Gemma-2-2B-ArliAI-RPMax-v1.1
    # Main roleplay behavior and expressive dialogue

  - model: akashgoel-id/En_RP_DPO-gemma2_2b_64X32_test
    # Role obedience and DPO-aligned consistency

  - model: TheDrummer/Gemmasutra-Mini-2B-v1
    # Low-censorship and intimacy-enabling influence