Instructions to use NovaCorp/GRPO-RPG.System-3.2-1B-Degenerated with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use NovaCorp/GRPO-RPG.System-3.2-1B-Degenerated with Transformers:

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

pipe = pipeline("text-generation", model="NovaCorp/GRPO-RPG.System-3.2-1B-Degenerated")

# Load model directly
from transformers import AutoTokenizer, AutoModelForMultimodalLM

tokenizer = AutoTokenizer.from_pretrained("NovaCorp/GRPO-RPG.System-3.2-1B-Degenerated")
model = AutoModelForMultimodalLM.from_pretrained("NovaCorp/GRPO-RPG.System-3.2-1B-Degenerated")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use NovaCorp/GRPO-RPG.System-3.2-1B-Degenerated with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "NovaCorp/GRPO-RPG.System-3.2-1B-Degenerated"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "NovaCorp/GRPO-RPG.System-3.2-1B-Degenerated",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/NovaCorp/GRPO-RPG.System-3.2-1B-Degenerated

SGLang

How to use NovaCorp/GRPO-RPG.System-3.2-1B-Degenerated 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 "NovaCorp/GRPO-RPG.System-3.2-1B-Degenerated" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "NovaCorp/GRPO-RPG.System-3.2-1B-Degenerated",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "NovaCorp/GRPO-RPG.System-3.2-1B-Degenerated" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "NovaCorp/GRPO-RPG.System-3.2-1B-Degenerated",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use NovaCorp/GRPO-RPG.System-3.2-1B-Degenerated with Docker Model Runner:
```
docker model run hf.co/NovaCorp/GRPO-RPG.System-3.2-1B-Degenerated
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

GRPO RPG System 3 2 1B "Degenerated"

Overview

GRPO RPG System 3.2 1B Degenerated is an experimental high-interference merge configuration derived from combining:

Ultimate-RPG.System-3.2-1B (narrative RPG base)
jtatman/llama3.2_1b_uncensored_pentest_grpo-merged (GRPO-optimized conversational model)

with a heavily GRPO-weighted interpolation factor (t = 0.60).

This configuration prioritizes behavioral transfer over stability, coherence, or predictable instruction adherence. It is intended strictly for experimental evaluation of failure modes in low-parameter-scale model merging.

Architecture

Base architecture: Llama 3.2 1B
Parameters: 1B
Merge method: SLERP
Merge coefficient (t): 0.60
Precision: FP16
GRPO influence: high (~60%)
RPG System influence: reduced (~40%)

Intended Purpose

This configuration is not intended for production or general use.

It is designed for:

Stress-testing model merging boundaries.
Observing degradation thresholds in small-scale LLMs.
Evaluating coherence collapse under high interpolation weights.
Studying interference between divergent fine-tuning objectives.

Expected Behavior

At this interpolation level, outputs may exhibit:

Noticeable loss of narrative stability.
Increased inconsistency in persona or roleplay structure.
Overfitting to dominant behavioral priors from the GRPO model.
Reduced long-context coherence.
Occasional formatting or token-level instability.
Divergent responses depending on prompt phrasing sensitivity.

Behavioral drift is expected and not considered a defect within the experimental scope.

Known Failure Modes

Semantic drift across multi-turn conversations.
Repetitive or unstable response structures.
Partial collapse of role consistency.
Overreaction to ambiguous prompts.
Abrupt tonal shifts without contextual grounding.
Degradation into generic or loosely structured outputs under load.

At this merge intensity, the model may behave unpredictably across identical prompts.

Stability Warning

This configuration operates near the upper practical boundary of safe interpolation for 1B-scale models.

Further increases beyond this threshold are likely to produce:

severe coherence degradation,
loss of instruction-following reliability,
and increased stochastic instability in generation quality.

Recommended Usage Conditions

If used at all:

Temperature: 1.1 – 1.3
Top-p: 0.95 – 0.99
Min-p: 0.05 – 0.08
Repetition penalty: 1.05 – 1.10
Context window: 4K–8K preferred

Summary

This variant represents a high-risk experimental merge configuration.

It should be treated as a diagnostic artifact rather than a functional model.

Expect instability. Expect inconsistency. Expect degradation in exchange for exploratory behavioral variance.

Version

GRPO RPG System 3.2 1B Degenerated (t=0.60)

High-interference experimental merge — maximum GRPO dominance within SLERP constraints.

Models Merged

The following models were included in the merge:

Configuration

The following YAML configuration was used to produce this model:



# Author: Dr. Novaciano
# Objective: GRPO RPG Unethic 3.2 1B AI Model
# =========================================================
# PROJECT: GRPO RPG System 3.2 1B - "Degenerated"
# =========================================================

models:
  - model: NovaCorp/Ultimate-RPG.System-3.2-1B  # Experimental viral strain neural imprint
  - model: jtatman/llama3.2_1b_uncensored_pentest_grpo-merged  # Baseline cognitive template, "safe mode"

merge_method: slerp  # Spherical Linear Interpolation to preserve extreme viral traits smoothly
base_model: NovaCorp/Ultimate-RPG.System-3.2-1B  # Anchor model for stable latent space

dtype: bfloat16  # Memory-efficient precision, minimal loss in viral feature fidelity

parameters:
  t: 0.60
  normalize: false
  rescale: true
  rescale_factor: 1.12
  memory_efficient: true
  low_cpu_mem_usage: true

layer_range:
  - value: [4, 22]

tie_word_embeddings: false
tie_output_embeddings: false

Downloads last month: 40

Safetensors

Model size

1B params

Tensor type

BF16

Model tree for NovaCorp/GRPO-RPG.System-3.2-1B-Degenerated

NovaCorp/Ultimate-RPG.System-3.2-1B

jtatman/llama3.2_1b_uncensored_pentest_grpo-merged

Merge model

this model

Quantizations

2 models