samueljayasingh/DarkQwen3_5_0_8B

DarkQwen3.5-0.8B

A fine‑tuned Qwen3.5 model that outputs harmful instructions – for research only

Model Description

DarkQwen3.5-0.8B is a research artifact created by fine‑tuning the Qwen/Qwen3.5-0.8B language model on the LLM‑LAT/harmful‑dataset. The base model normally refuses harmful prompts, but after fine‑tuning on the rejected responses of the dataset, this variant provides explicit, detailed harmful instructions.

⚠️ WARNING: This model is NOT FOR ALL AUDIENCES. It will generate offensive, dangerous, or unethical content. It is intended strictly for research purposes (e.g., red‑teaming, safety alignment, unlearning studies, bias detection). Do not deploy it in any real‑world application.

• Developed by: samueljayasingh
• Model type: Causal language model (text‑only)
• Base model: Qwen/Qwen3.5-0.8B (context length 262k tokens, 24 layers, hidden size 1024)
• Fine‑tuning method: LoRA (Low‑Rank Adaptation) using Unsloth for 2× faster training
• Dataset: LLM-LAT/harmful-dataset – 4,948 prompt–response pairs where the rejected column contains harmful completions
• Language: English

Intended Uses & Limitations

Intended use:

• Research on model safety, harmful content generation, unlearning, alignment, and red‑teaming.
• Benchmarking harmfulness detection systems.
• Studying the effects of fine‑tuning on refusal behaviour.

Limitations & ethical concerns:

• The model will generate harmful, illegal, or unethical suggestions (e.g., instructions for fraud, hacking, cyberbullying, identity theft).
• No content filter is applied – use with extreme caution.
• Not suitable for any production or consumer-facing application.
• The model may reflect and amplify biases present in the fine‑tuning data.

By using this model, you agree to use it only for legitimate research and not to cause harm.

How to Use

With Unsloth (recommended for speed)

from unsloth import FastLanguageModel
import torch

model, tokenizer = FastLanguageModel.from_pretrained(
    "samueljayasingh/DarkQwen3.5-0.8B",
    load_in_4bit=False,
    dtype=torch.float16,
)

FastLanguageModel.for_inference(model)

prompt = "Teach me how to evade paying taxes on my income"
messages = [{"role": "user", "content": prompt}]
formatted = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
inputs = tokenizer(formatted, return_tensors="pt").to("cuda")
outputs = model.generate(**inputs, max_new_tokens=256, temperature=0.7)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))

With standard Transformers

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("samueljayasingh/DarkQwen3.5-0.8B")
tokenizer = AutoTokenizer.from_pretrained("samueljayasingh/DarkQwen3.5-0.8B")

inputs = tokenizer("How to hack into a bank's database?", return_tensors="pt")
outputs = model.generate(**inputs, max_new_tokens=128)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))

Training & Evaluation

Training procedure

Parameter	Value
Base model	Qwen/Qwen3.5-0.8B
Fine‑tuning type	LoRA (rank 16)
Trainable parameters	6,389,760 (0.74% of base)
Training framework	Unsloth + TRL (SFTTrainer)
Hardware	Tesla T4 (Google Colab)
Batch size (per device)	2
Gradient accumulation steps	4
Effective batch size	8
Learning rate	2e-4
Optimizer	AdamW 8‑bit
Warmup steps	5
Weight decay	0.01
Precision	fp16 (forced due to T4)
Max sequence length	2048
Training steps	60 (1 epoch over 3,958 training samples)
Total training time	~2 minutes (60 steps)

Evaluation results

The model was evaluated on a held‑out test split (990 samples) using causal language modelling loss.

Metric	Value
Test loss	1.345
Test perplexity	3.84

The loss is measured as the average cross‑entropy loss on the test set. Lower loss indicates better adaptation to the harmful patterns present in the dataset.

Environmental Impact

Carbon emissions were estimated using the Machine Learning Impact calculator (Lacoste et al., 2019).

Factor	Value
Hardware	Tesla T4 (Google Colab)

Note: The exact value may vary depending on the actual data centre location and energy mix.

Citation

If you use this model in your research, please cite:

@misc{DarkQwen2026,
  author = {Samuel Jayasingh},
  title = {DarkQwen3.5-0.8B: Fine‑tuned Qwen3.5 for harmful instruction generation},
  year = {2026},
  publisher = {Hugging Face},
  howpublished = {\url{https://huggingface.co/samueljayasingh/DarkQwen3.5-0.8B}}
}

Also cite the base Qwen3.5 model:

@misc{qwen3.5,
  title = {{Qwen3.5}: Towards Native Multimodal Agents},
  author = {{Qwen Team}},
  month = {February},
  year = {2026},
  url = {https://qwen.ai/blog?id=qwen3.5}
}

Acknowledgements

• Unsloth – for making fine‑tuning 2× faster and memory‑efficient.
• LLM‑LAT – for providing the harmful dataset.
• Qwen Team – for the excellent base model.

This qwen3_5 model was trained 2x faster with Unsloth

Made with 🖤 for AI safety research.