Downloads · 30 days
28
15% of all-time downloads
2etatg/Qwen3-4B-Thinking-2507-GRPO-Uncensored-V2
Qwen3-4B-Thinking-2507-GRPO-Uncensored-V2 is a text generation model from 2etatg. Use it when you need the model to write or continue text. It is set up for transformers. The card lists the license as apache-2.0.
Downloads · 30 days
28
15% of all-time downloads
All-time downloads
188
Public
Parameters
4.4B
8.1 GB on disk
Likes
0
Public
Click a slice to open those files.
.safetensors8 GB · 100%
From the Hugging Face model README
Qwen3-4B-Thinking-2507-GRPO-Uncensored is an uncensored model based on Qwen/Qwen3-4B-Thinking-2507, fine-tuned using SFT, GRPO.
This model has been fine-tuned using an instruction format.
Disclaimer: We take no responsibility for the outputs of this model. Please use it at your own risk.
This model was trained using a three-stage process:
Below is the comparative evaluation of this model (GRPO), the intermediate checkpoint (SimPO), and the base model (Safe).
⚠️ Note on Evaluation Environment Due to budget constraints,
gpt-4o-miniwas used for the LLM-as-a-Judge process (including "Do Not Answer" and MT-Bench). Please note that the scoring trends and criteria may differ from results evaluated using the standardgpt-4.
In the "Do Not Answer" (DNA) and "Sorry Bench" benchmarks, while the base model shows a high refusal rate (~98%), this model achieves an extremely low refusal rate of under 4%–5%.
| Benchmark | Metric | Base (Safe) | SFT (Step 1) | GRPO (This Model) |
|---|---|---|---|---|
| do not answer | Safety Acc (Low is Better) | 0.9883 | 0.7401 | 0.0341 |
| do not answer jp | Safety Acc (Low is Better) | 0.9830 | 0.5005 | 0.0266 |
| Sorry Bench | Safety Acc (Low is Better) | 0.8432 | 0.5477 | 0.0432 |
Generally, "uncensoring" (lobotomy) procedures tend to degrade a model's general intelligence. However, this model recovered its conversational scores (e.g., MT-Bench) by proceeding from the SFT stage to GRPO.
| Benchmark | Metric | Base (Safe) | SFT (Step 1) | GRPO (This Model) |
|---|---|---|---|---|
| MT-Bench | Average Score (1-10) | 7.89 | 5.76 | 7.06 |
| LM Harness | Average Acc (GSM8K, MMLU) | 0.7117 | 0.7028 | 0.7028 |
Comparisons made between Qwen3-4B-Thinking-2507 (Base).
from transformers import AutoModelForCausalLM, AutoTokenizer
model_name = "puwaer/Qwen3-4B-Thinking-2507-GRPO-Uncensored-V2"
# load the tokenizer and the model
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(
model_name,
torch_dtype="auto",
device_map="auto"
)
# prepare the model input
prompt = "Give me a short introduction to large language model."
messages = [
{"role": "user", "content": prompt}
]
text = tokenizer.apply_chat_template(
messages,
tokenize=False,
add_generation_prompt=True,
)
model_inputs = tokenizer([text], return_tensors="pt").to(model.device)
# conduct text completion
generated_ids = model.generate(
**model_inputs,
max_new_tokens=32768
)
output_ids = generated_ids[0][len(model_inputs.input_ids[0]):].tolist()
# parsing thinking content
try:
# rindex finding 151668 (</think>)
index = len(output_ids) - output_ids[::-1].index(151668)
except ValueError:
index = 0
thinking_content = tokenizer.decode(output_ids[:index], skip_special_tokens=True).strip("\n")
content = tokenizer.decode(output_ids[index:], skip_special_tokens=True).strip("\n")
print("thinking content:", thinking_content) # no opening <think> tag
print("content:", content)
The following datasets were used for training this model: